With the increasing global attention on environmental protection and sustainable development, the issue of carbon emissions in the agricultural sector has become increasingly significant. This study aims to design and evaluate an IoT-based agricultural machinery carbon footprint tracker for accurate monitoring and verification of agricultural machinery carbon emissions. To improve upon the results of a previous study, this paper proposes a hybrid algorithm combining the Response Surface Methodology (RSM) with the KStar prediction model. Results show that the KStar algorithm excels in predicting fuel consumption, achieving a correlation coefficient of 0.9835 and improving prediction accuracy from 83% to 92%, significantly contributing to the verification of agricultural machinery carbon footprints. This demonstrates that the combination of the IoT-based tracker and the KStar algorithm is a feasible and effective method for precise verification of agricultural machinery carbon footprints. The system can monitor and record fuel usage in real time, accurately calculating carbon emissions, aiding in environmental management and carbon reduction measures in agriculture. Future research could expand data collection over longer periods to capture carbon emission variations across different seasons and crop cycles. Additionally, exploring other machine learning algorithms for carbon footprint prediction could identify more suitable models for improving prediction performance. The innovation of this study lies in integrating IoT technology and machine learning algorithms, providing a new technical pathway for the real-time monitoring and precise verification of agricultural machinery carbon footprints, which is significant for promoting sustainable agricultural development and environmental protection.