Abstract: Abstract: Soil degradation and non-point source pollution have posed a great challenge to sustainable agriculture, due mainly to the excessive application of chemical fertilizers. Among them, High-standard Farmland Construction (HSFC) can be the paramount component of the national strategy for food security and environmental protection. Taking the HSFC in 2013 as a quasi-natural experiment, this study aims to clarify the impacts of the HSFC policies on fertilizer reduction at the county level. Firstly, the datasets were divided into the experimental and control groups with/without HSFC policies. A t-value test was carried out to determine the mean differences in the characteristics between the treatment and control groups. A Differences-in-Differences (DID) estimator was used to assess the effects, using the panel data of 282 counties in 3 major grain-producing provinces in the Middle and Lower Reaches of the Yangtze River from 2007 to 2017. An event analysis was then constructed to verify the reliability and robustness, particularly for the pre-event parallel trend and joint hypotheses tests from 2007 to 2012, and 2014 to 2017. Secondly, a series of robustness tests were performed, including the measurement of the dependent variable and regression model, considering the lag in the effect of the policy, and excluding the impact of other similar policies. Finally, the agricultural mechanization and the grain production specialization were selected to reveal the conversion after the HSFC in the study area. The mediation effect model was employed to test the mechanism of agricultural mechanization, and the grain production specialization between the HSFC and chemical fertilizer reduction, together with the interaction between the agricultural mechanization and the grain production specialization. After that, the heterogeneity test was performed to determine the differences in the policy impact in the regions. The results showed that: 1) The density of fertilizer application in the treatment groups was higher than that in the control from 2007 to 2013, whereas, it was reversed after the HSFC. The t-value test showed that there was a significant difference in the mean of core variables between the treatment and control groups. 2) The HSFC policies exerted a significant reduction effect on the chemical fertilizer application in the exampled grain dominant counties from 2007 to 2017, with an average of 5.1% decrease in the fertilizer application. There was a consistent trend between the treatment and the control groups before the HSFC, indicating a basically stable reduction. By contrast, there was a slight reduction after excluding the impact of the "double reduction policy". 3) The impacts of HSFC on the chemical fertilizer application depended mainly on agricultural mechanization and grain production specialization. There was a gradual increase in fertilizer reduction in the HSFC, with the development of agricultural mechanization and grain production specialization. Furthermore, mechanization greatly enhanced production specialization and the HSFC impact on fertilizer reduction. 4) The heterogeneity analysis showed that the HSFC effects varied in the fertilizer application. There was a strong reduction in the fertilizer application in the plains and shallow hills areas, whereas, a higher level of fertilizer application was found in the counties of central provinces. Therefore, agricultural mechanization and specialization can be the overall measures to achieve agricultural development and ecological protection in the process of HSFC.