The renewable energy accommodation obligation weight system is a key institutional guarantee for promoting the integration of wind and photovoltaic power. However，the current weight allocation mechanism adopts a uniform standard，failing to fully account for the disparities in regional resource endowments and development foundations，which can easily lead to inter-provincial fairness concerns and efficiency losses in resource allocation. To address this，this paper proposes a provincial wind and solar accommodation obligation weight allocation method based on differentiated grouping. Firstly，an accommodation capacity evaluation system encompassing multi-dimensional indicators such as resource conditions，grid structure，and load characteristics is constructed. Hierarchical clustering is applied to categorize provinces into clusters that are “homogeneous within groups and heterogeneous between groups，”allowing for precise identification of differences in accommodation potential. Secondly，based on historical data and a combination weighting method，differentiated allocation between groups and initial allocation within groups are performed. Subsequently，an improved zero-sum gains data envelopment analysis model is introduced to optimize the efficiency of weight allocation within groups. Finally，the Gini coefficient and Theil index are employed to quantitatively evaluate the fairness of the allocation results. The results show that compared to the traditional allocation scheme，this method reduces the two aforementioned fairness indicators by 7.32% and 6.15% respectively，effectively enhancing the rationality and fairness of responsibility sharing for wind and solar power accommodation among provinces.