Agricultural carbon emissions significantly contribute to global greenhouse gases. Enhancing green and low-carbon agricultural practices is crucial for China’s high-quality economic progression and achieving its “carbon peaking and carbon neutrality” objectives. This study focuses on agriculture’s ecological role, incorporating 18 primary carbon sources across agricultural materials, soil, water fields, and animal husbandry into an evaluative framework. It assesses the total agricultural carbon emissions in 31 Chinese provinces from 2000 to 2021. Employing the STIRPAT environmental pressure model, the paper investigates the determinants of China’s agricultural carbon emissions. Additionally, it utilizes the BP neural network model for forecasting emission peak trends under various development scenarios and validates these predictions through the Geographically and Temporally Weighted Regression (GTWR) model, among other methods. The findings reveal a reverse U-shaped pattern in China’s total agricultural carbon emissions over the study period, marked by initial growth followed by a decline and significant regional variations. The primary drivers of these emissions are the agricultural population, per capita agricultural GDP, and agricultural technology level. Under green development initiatives, China’s agricultural sector is projected to achieve its “peak CO₂ emission” goal by around 2030, with minimal peak variations. This research offers valuable insights into Chinese agriculture’s carbon sequestration capabilities within the context of carbon peak and neutrality goals. It guides governmental agencies in devising flexible, precise, and moderate agricultural carbon sink strategies, enhancing regional agricultural collaborations, and promoting pollution and carbon reduction in China’s agriculture towards realizing its “carbon peaking and carbon neutrality” ambition.