Abstract: Using the observation data of daily maximum and minimum temperature in Changxing island and Jiaoliu island weather stations from August 2015 to July 2017,the 2 m fine ECMWF grid temperature forecast value,the Japanese FSFE02 (24 h ground situation forecast) and FSFE03 (36 h ground situation forecast) were verified.The results show that on the basis of the historical regression statistical test method,the forecasted 24 hours maximum and minimum temperatures at 2 m using the ECMWF fine grid model have significant forecast effects and pass the 0.05 confidence test.A close correlational relationship between the prediction and observation for each month is obtained.Making use of the difference of the 2 m temperature between the observation and the forecast value from ECMWF fine mesh the day before in Changxing island,integrating the statistical correction of the 2 m temperature forecast values from the ECMWF fine grid,the temperature for the next 24 hours at Changxing island stations is forecasted.The temperature forecast of Jiaoliu island station is produced based on that of Changxing island station.On the basis of the statistical analysis,the temperature difference between Jiaoliu island and Changxing island station is closely related to the ground situation and wind field.Based on the corrections of different ground conditions and wind field,the temperature at Jiaoliu island station is forecasted.Applying development function of the Matlab computer language to extract the maximal and minimal temperature at 2 m from the ECMWF fine mesh and to input their real values at Changxing and Jiaoliu island stations,the maximum and minimum temperatures in the next 24 hours are automatically forecasted.Besides,a visual forecast working platform is built to realize the automation of villages and towns temperature forecast.