Abstract:

[Objective] Smart agriculture serves not only as a critical measure for driving agricultural and rural modernization, but also as a core component of new quality productive forces in agriculture. In recent years, China's smart agriculture has achieved remarkable development. However, it still faces practical challenges in policy implementation, technological application, and industrial coordination, which restricts the high-quality development of agriculture. Against this background, quantitative analysis of the central-level smart agriculture policy system was conducted from the crucial link of policy design. Grounded in practical dilemmas of smart agriculture development and top-level guiding requirements, policy logic and inherent deficiencies were clarified via scientific analytical approaches.The aim of this research is to identify the direction of policy optimization, so as to provide practical references for promoting the quality and efficiency improvement of smart agriculture and advancing the construction of a strong agricultural country. [Methods] From the perspective of policy tools, first, a three-dimensional analytical framework of "tools-topics-objectives" was constructed, which served as the core analytical logic of the overall research. A total of 240 central-level policy documents concerning smart agriculture issued from 2013 to 2025 were selected as the research sample. Systematic coding and quantitative analysis of all collected policy texts were implemented through content analysis with the assistance of NVivo14 qualitative analysis software. To further explore the latent thematic structures of China's smart agriculture policy texts, accurately identify the focal directions of policy content and their long-term evolutionary characteristics, the Latent Dirichlet Allocation (LDA) topic model was further adopted to realize automatic text topic analysis. Through the above integrated research methods and analytical pathways, core characteristics of China's smart agriculture policies were systematically summarized, and targeted optimization strategies were put forward. [Results and Discussions] First, the analysis identified four distinct evolutionary stages: the Policy Exploration Period (2013－2015), the Pilot Demonstration Period (2016－2018), the Strategic Promotion Period (2019－2023), and the Comprehensive Deepening Period (2024 to present). This progression reflected a shift from conceptual introduction and top-level design to concrete technology application, strategic scaling, and finally, systematic institutionalization. Second, in terms of policy instrument deployment, it was found that supply-side, environmental, and demand-side tools were used in a relatively balanced overall proportion. However, significant structural imbalances existed within each category. Supply-side tools were heavily skewed towards infrastructure construction and scientific research support, with relatively less emphasis on direct financial input and talent cultivation. Environmental tools were dominated by publicity and guidance, while financial support mechanisms and standard-setting were underutilized. On the demand side, policy relied heavily on pilot demonstrations, with weaker emphasis on fostering cooperation, industry-research integration, and broader social participation. Thirdly, the evolution of policy objectives exhibited an obvious phased and progressive feature, whereas its internal structure still needed optimization. On the whole, national smart agricultural policies covered five core fields, with the digital and intelligent transformation of production and operation as the primary objective. Technological innovation and data services acted as crucial supporting pillars. By contrast, two guarantee-oriented objectives, namely talent team development and institutional standard improvement, accounted for merely 10.91% and 10.13%, respectively, resulting in a prominent imbalance in the objective system. Furthermore, policy development could be divided into four sequential stages: Initial conceptual guidance and foundational planning, pilot-oriented practical implementation, systematic policy construction, and in-depth quality and efficiency upgrading, forming a clear progressive trajectory. Though digital transformation had long served as the central focus across all stages, the development of talents and institutional mechanisms had witnessed gradual yet relatively slow progress, constantly lagging behind the advancement of core policy objectives. Fourthly, the LDA topic model extracted four core policy themes: Digital transformation and intelligent upgrading, industrial integration and business model innovation, talent cultivation and entrepreneurial support, and technical equipment and intelligent application. A striking finding was the extreme concentration of policy attention, with the digital transformation and intelligent upgrading theme accounting for 57.92% of the thematic focus, significantly overshadowing the other three themes combined, the matching degree between policy themes and policy instruments needs to be improved. [Conclusions] While China has developed a comprehensive policy framework for smart agriculture, its effectiveness is hampered by structural imbalances in instrument use, an over-concentration on a single policy theme, and an inconsistent coordination of policy objectives. To address these challenges and foster a more robust and sustainable development path, three key optimization strategies are proposed. First, promoting the balanced deployment of policy tools by following a tripartite approach: upgrading supply-side tools, empowering environmental-side tools, and driving demand-side tools, so as to enhance synergy and complementarity among different tool types. Second, optimizing the allocation of attention to policy themes and enhance the alignment between policy instruments and policy themes, avoiding over-reliance on individual fields and under-support for key sectors. Third, advancing the coordinated optimization of policy goals, and improving the systematic mechanism led by standards, guaranteed by institutions, and supported by talent teams. Implementing these strategies will be crucial for transforming China's smart agriculture policy system from one focused on quantity and foundational building to one capable of driving high-quality, balanced, and innovative development in the years to come.

Key words: smart agriculture, policy evolution, policy tools, policy topics, policy objectives, Latent Dirichlet Allocation (LDA) topic model