Precise forecast of precipitation in convective scales is very important. Since the initial conditions of the forecasts can be improved using convective scale data assimilation, therefore, data assimilation can have a positive impact on precipitation prediction. In this study, the effect of cyclic data assimilation using 3DVAR method and WRF model on convective scale precipitation forecasts with 2.5 km spatial resolution is investigated. The one month period (December 2013) has been selected for numerical simulations and model runs have been conducted in a select area in western and northwestern of Iran. The used boundary conditions data are extracted from ECMWF model predictions. Also, various types of conventional observations were obtained from ECMWF center for data assimilation process and with a domain specific background error statistics that it is produced by CV5 method in WRF-DA system. Overall, the predicted precipitation data for many low rainfall thresholds showed higher verification scores, although this improvement was not observed at rainfall thresholds greater than 10 mm.