In this study, the WRF (Weather Research and Forecasting) model was implemented to predict the atmospheric conditions, particularly the precipitable water vapor (PWV) in the North of Chile. Its performance was evaluated over the ALMA (Atacama Large Millimeter/submillimeter Array) site. Five WRF configurations with different physical options for boundary layer, soil model and microphysics were compared with observations from a radiometer and a weather station from April to December 2007. The results show that all the simulations overestimate PWV values, particularly in summer months. In addition, the microphysics parameterization changes do not notably affect the forecast, observing improved results with the soil model Noah. The errors were smallest with the YSU-Noah configuration, suggesting that it is appropriate to be used in operational forecasting of PWV in ALMA.