Based on the fixed data of the national rural observation points from 2009 to 2015, this paper empirically tests the effect of financial support for agriculture, especially agricultural subsidies, by means of income liquidity matrix and panel quantile regression. The results show that the financial subsidies do not significantly improve the income flow of rural residents. Although it has a positive impact on low-income farmers, it will not change the overall rural income distribution pattern. However, financial subsidies can help alleviate the absolute poverty in rural areas and effectively increase the income of low-income groups. However, compared with other fiscal transfer expenditures, the role of agricultural subsidies in poverty reduction does not have an advantage. In order to improve the accuracy of the subsidy policy, this paper proposes to adjust the existing subsidy methods, that is, instead of simply linking the GSP subsidy with the land area, the whole subsidy is divided into two categories: one is to integrate with the rural minimum living insurance fund to support the rural vulnerable groups; the other is to focus on the appropriate scale of the operating subjects and support through supporting agricultural insurance and guarantee the construction of agricultural socialized service system.