The spatial distribution of precipitation is a key data for the prevention and management of extreme events that threaten the Assaka watershed. This area is characterized by a scarcity of climatological data, an unevenly distributed rainfall observation network and low density. However, spatial interpolation methods of point precipitation measurements could overcome these aspects. For this reason, this research consists in determining the most adequate method in terms of efficiency and practical use in order to accurately map the maximum daily precipitation for a period of 30 years (1990 -2020). In this context four interpolation techniques (Thiessen polygons, inverse distance weighting, ordinary kriging and linear regression) were applied in a GIS environment. The cross-validation allows to evaluate the global performance of each method using statistical indicators (RMSE, MAE) as well as adjustment diagrams between observed and predicted values. Indeed, this analysis has allowed to qualify the method of multiple linear regression (MLR), as the best interpolator (RMSE=1.67mm and MEA=1.40mm). These results are judged by the fact that this technique integrates geographical factors (topography, latitude, proximity to the ocean) related to the formation of precipitation in the study area. Other methods are considered unsuitable in this anisotropic environment where the density of observation points is very low. These results constitute exploitable approaches by scientists and decision-makers in the prevention and management of extreme events (floods, landslides, water erosion) as well as land management (water resources, agriculture and environment).