Based on the daily 2m maximum and minimum temperature forecast data of SCMOC, SNWFD and SPCO models in 2016-2018 and actual temperature data of stations, this paper analyzes and discusses the forecast performance of refined and revised products in Sichuan. The results show that: (1) Among the three models, the prediction capabilities of SPCO and SNWFD for the max/min temperatures are comparable, and both are higher than SCMOC. (2) The forecasting effect in the eastern part of Sichuan is better than that in the western region, and the forecast error is the greatest in the western Sichuan Plateau and Liangshan Prefecture. (3) For the seasons, the forecasting effect is better in summer and autumn than that in the spring and winter. (4) The forecast error increases with the increase of the forecasting time. (5) The forecast of the minimum temperature in Sichuan is better than that of the maximum temperature.