The main goal of this research was to evaluate the potential of the HYDRUS-1D numerical model for estimating the soil moisture (θ) at different depths, actual crop evapotranspiration (ETa) and its components (crop transpiration, Ta and soil evaporation, Ea) as well as the deep percolation (DP) of irrigated winter wheat under different water managements in the semi-arid region of Tensift-basin (central Morocco). The HYDRUS-1D simulations were performed at daily time step during the two growing seasons: 2002/2003 and 2015/2016. The model was firstly calibrated based on one field "denoted F1" data during the 2002/2003 cropping season by using the Levenberg-Marquardt method implemented in HYDRUS-1D model for optimizing various parameters of Van Genuchten equation that provide the minimum difference between measured and simulated soil moisture at four layers of soil (0-5, 5-10, 10-20, 20-30, 30-50 cm). Afterwards, the model validation was done based on the data from four fields of wheat: two fields "denoted F2 and F3" during the 2002/2003 and two other fields "denoted F4 and F5" during the 2015/2016 cropping season. All fields were irrigated with flooding system except the field F5 where drip irrigation was undertaken. In-situ measurements of θ was carried out using Time Domain Reflectometry (TDR) and gravimetric method ETa was measured by the Eddy Covariance system Ta and Ea were monitored using a lysimeter in F5 field. The results showed that the HYDRUS-1D model simulates the θ, ETa, Ta and Ea reasonably well. Additionally, the evaluation of the irrigation system on DP losses was investigated by comparing the simulation results over flood (F4) and drip (F5) irrigated fields. It was found that about 56% and 20% of seasonal supplied water were lost by DP in F4 and F5 sites, respectively. Such unexpected high amount of DP taking place in F5 field is due to the improper use of the drip irrigation system..