The United Nations 2030 Agenda for Sustainable Development (UN SDGs) frames sustainable agriculture as a core pathway to advance zero hunger (SDG 2) and protect life on land (SDG 15). As critical platforms for the green transition of agriculture, innovative perception among new-type agricultural business entities inside national modern agricultural industrial parks constitutes a prerequisite cognitive factor shaping sustainable farming practices. Taking Anhua National Modern Agricultural Industrial Park—China’s sole industrial park specialized in dark tea—as the research case, this study develops a four-dimensional measurement framework for innovative perception covering innovation cognition, motivation, demand, and intention, grounded in 363 field micro-survey questionnaires. Factor analysis is adopted to quantify township-level composite innovative perception scores, and the Geodetector model is integrated to detect driving factors as well as their interactive coupling effects behind spatial disparities. The results indicate the following: (1) Marked spatial heterogeneity characterizes innovative perception levels of new-type agricultural business entities across the park, which maintain an overall high standard. The core zone registers an average score of 4.209 versus 3.726 for the demonstration zone; innovative perception in the core zone is statistically higher, presenting a typical core–periphery gradient pattern across townships. (2) Individual endogenous attributes, including age, educational attainment, and farming experience, show strong explanatory capacity. The interaction between age and innovation applicability yields a q-statistic of 0.961, the strongest pairwise spatial association observed within the present sample; nevertheless, q-statistics from the Geographical Detector are sample-specific descriptive metrics for spatially stratified explained variance and do not represent causal effect sizes. This research uncovers the core–periphery polarization rule and multi-factor spatial coupling patterns underlying innovative perception of operators within specialized agricultural industrial zones. The empirical micro-evidence generated in this work informs the design of tiered, differentiated governance policies aimed at mitigating intra-park spatial disparities, delivering empirical insights into the cognitive foundations of the park’s green transition. This analysis targets innovative perception as the cognitive prerequisite for sustainable farming practices instead of direct sustainability outcomes; empirical tracing of the perception-to-behavior-to-outcome causal chain would require longitudinal outcome-focused datasets, which constitutes a key avenue for future inquiry.

Driven by the goal of carbon neutrality, low-carbon development in township spaces is essential for sustainable urban–rural growth. This paper employs a carbon accounting methodology, taking Fuqiushan Town in the Dongting Lake Ecological Economic Zone as a case study to develop a detailed carbon measurement inventory at the township scale. Using spatial analysis techniques, it synthesizes multi-source data—including land use, agricultural inputs, and population—to estimate emissions from key sources such as crop cultivation, livestock and poultry breeding, industrial production, and residential activities. The study also evaluates the carbon sequestration capacity of sinks such as woodlands and water bodies, enabling the spatial visualization of both carbon emissions and carbon sinks. Key findings include: (1) Fuqiushan Town exhibits a carbon emission profile characterized by “industrial activities as the primary source, supplemented by agriculture, with additional contributions from residential and transportation sectors,” while forested areas and water bodies serve as core carbon sink zones. (2) An innovative multidimensional indicator system for low-carbon development efficiency was established, consisting of the Low-Carbon Development Efficiency Index in Production, the Daily Life Carbon Responsibility Efficiency Index, and the Ecological Carbon Sink Efficiency Index, which together form a Comprehensive Efficiency Index for Low-Carbon Development. (3) Analysis reveals significant spatial coupling relationships and efficiency differentiation patterns among carbon emissions, industrial structure, energy dependence, and ecological background. Based on dominant carbon emission types, low-carbon efficiency thresholds, and spatial factor interactions, the 17 villages and one forest farm in the township are classified into five zones: “Industrial High-Carbon Transition Zone,” “Agricultural Pollution Reduction and Carbon Emission Reduction Synergy Zone,” “Ecological Low-Carbon Conservation Zone,” “Human Settlements Balanced Development Zone,” and “Ecological Core Zone.” Tailored low-carbon spatial planning strategies for material resources are proposed for each zone. These results offer quantitative support and spatially targeted insights for low-carbon spatial planning in ecologically sensitive townships, contributing to the achievement of objectives such as “carbon reduction and sink increase” and “rural revitalization.”