Abstract: Based on the 1km and 5km multi-source melting qualification point data and ground observation station data in Sichuan,the prediction accuracy,average absolute error,root mean square error and Alpha Index (AI) statistics were adopted,the quality of the actual data of the multi-source melting qualification points was evaluated by selecting two high temperature weather processes in Sichuan in the summer of 2020. The results show that the adjacent interpolation method is better than bilinear interpolation for multi-source fusion qualified point data. The large error areas are mainly located in the high-altitude areas,such as the western Sichuan Plateau,the Panxi area and the mountain week of the basin. AI index is close to 0,and the actual data of the qualified point of multi-source melting has no random error and is closer to the ideal value. The applicability of the 1km resolution melting qualified point data in Sichuan is better than 5km,the accuracy of error ≤2℃ can reach 98%,and the root mean square error is less than 1.