Abstract: The statistical and dynamical methods are used to interpret the mesoscale NWP products produced by the Beijing Area Mesoscale NWP System (which has been developed under a joint project of the National Meteorological Center and the Beijing Meteorological Bureau) in terms of the local weather elements and air pollution. First, the Kalman filtering method is applied to the mesoscale NWP products, generating the 3 h interval wind, temperature, and the daily maximum /minimum temperature forecasts over nine automatic weather stations in the Beijing area. The results of verification indicate that the forecasts of wind and temperature with Kalman filtering are much better than direct model outputs. Second, a simple dynamical air pollution model is connected to the Mesoscale NWP system, making daily air pollution forecasts. For the SO₂, NO₂ and CO, the predicted pollution levels are correct on more than 65% days during summer, fall and winter; for RSP, the forecasts are right on more than 65% days during fall and winter, but only on 50% days during summer. Both methods have been put into operational use.