Finemesh grid element forecast is important service in the China Meteorological Administration,and also the future development direction of weather forecast. How to effectively provide forecast products to users is the final step of finemesh grid point forecast. This article introduces a finemesh grid point forecast and display system in Shaanxi based on the BaiduMap application programming interface. This system uses the dynamical bias correction method to improve precipitation forecast,based on the hourly fusion precipitation data of CMORPH (NOAA Climate Prediction Center Morphing Method) satellites and gauged rainfall from about 30 000 automatic weather stations. At the same time, the moving linear regression method is used to improve temperature prediction. It can get geographic information of any position and display 240hour meteorological element forecast in realtime for the corresponding grids. Through calculating the forecast values with the finemesh grid value for 98 station spots in Shaanxi for 3 hours, the errors between the precipitation forecast, temperature forecast and the observed values are analyzed in the last 24 hours in realtime for users to anticipate possible errors in future to provide the daily element forecast of 98 stations in the next 168 hours. Different from the previous database background support, the system reads the 35 GB gridpoint forecast data directly into memory at one time, and intercepts, so to solve the low efficiency problem of database searches and calls.