Scientific understanding of greenhouse gas emissions from farmland ecosystems is an essential prerequisite for implementing agricultural emission reduction and carbon reduction measures, and is of significant importance for promoting sustainable agricultural development and facilitating high-quality construction of farmlands. Taking Anhui Province as an example, this study employs the IPCC emission factor method to assess the level of greenhouse gas emissions from farmland ecosystems. The Logarithmic Mean Divisia Index (LMDI) method is used to analyze the driving mechanisms behind these emissions. Finally, based on the STIRPAT model and combined with scenario analysis methods, future emissions from farmland ecosystems are simulated and predicted. The results show that: 1) The total greenhouse gas emissions from the farmland ecosystem in Anhui Province are of an overall increasing trend, with CH₄ emissions from rice paddies contributing the most, accounting for 55.27% of the total. The spatial distribution shows higher emissions in central Anhui and lower emissions in northern and southern Anhui. 2) In Anhui Province, the intensity of emissions per unit sowing area is increasing, while the intensity of emissions per unit of agricultural output value is decreasing. Both types of emission intensity exhibit a "low in the north and high in the south" distribution pattern, meaning that areas in north of the Huai River have relatively lower emission intensity and variation, while areas in south of the Huai River exhibit the opposite trend. 3) The level of agricultural economy has a positive effect on the greenhouse gas emissions from the farmland ecosystem in Anhui Province, being the main factor contributing to the increase in emissions; agricultural production efficiency, agricultural production structure, and agricultural population size all show negative effects, with agricultural production efficiency and population size having a stronger inhibitory effect on emissions. 4) Under baseline, low-carbon, green development, and sustainable development scenarios, the greenhouse gas emissions from farmland ecosystems show an initial increase to a peak followed by a gradual decrease, with the baseline scenario peaking in 2028 and the other three scenarios achieving their peak targets by 2025. Under the extensive development scenario, the greenhouse gas emissions from farmland ecosystems are projected to increase continuously in the future, failing to achieve the peak target but with a gradually slowing rate of increase, indicating significantly potential for emission reduction in agricultural development. Anhui Province should strengthen control over CH₄ emissions from rice cultivation, take into account regional differences, and formulate agricultural emission reduction policies that are multi-faceted and tailored to local conditions.