[Objective] The dynamics of soil moistures of different soil types and its relation to meteorological factors were studied in the Three-North Shelterbelt System Project of Western Heilongjiang Province, to provide reference for soil moisture prediction.[Methods] Using miniature meteorological station, soil moisture and relevant meteorological factors were monitored in situ. Their relationships were analyzed using multiple regression and correlation analysis methods.[Results] (1) Soil moisture showed a decrease trend in growing season for all the monitored soils, of which Phaeozem, Chernozem behaved more obvious. The variation coefficient of soil moisture for all the 3 soil types decreased with the increase of soil depth. (2) The correlation analysis showed that the soil moisture content was negatively correlated with light intensity and atmospheric temperature, but positively correlated with the air humidity. The correlation coefficients with rainfall and wind speed were comparatively smaller. (3) The multiple regression analysis showed that daily soil moisture consumptions of Phaeozem, Chernozem could be explained by illumination intensity(X₁), humidity(X₂), wind speed(X₃) and air temperature(X₄).[Conclusion] Soil moisture was regulated comprehensively by meteorological factors. The established multiple regression models could be used to forecast the changes of soil moisture in upland soil.