The best results for spatial interpolation methods depend on the high dataset network density, and interpolation performance also depends on station density and the specific spatial variability (surface topography) of the climate variables. And because climate data are difficult to obtain from ground-based meteorological stations in the Western Desert of Egypt, our research depends on the NASA POWER database to obtain a relative humidity dataset for the period 2010- 2022, where many studies have confirmed and validated the extent to which NASA POWER data is consistent with ground station data and can depend on overcome missing data from weather station sites. And then aimed to an operational application for spatial interpolation methods of spatial distribution to Annual data relative humidity and select the best spatial interpolation models among three spatial Interpolation algorithms Geostatistic interpolation methods ( Ordinary kriging, Empirical Bias kriging Regression Prediction), and deterministic spatial interpolation methods (Thin Plate Smooth Spline). The paper shows the comparison Statistic assessment and cartography visualization that the best models were based on ordinary kriging with ME: 0.002, MS: 0.006, RMS:1.4, RMSS: 1.1, thin TPS_Spline Model the better performance with MS: 0.001, RMS: 1.7, EBKRP Performance is not big different about other methods so its Statistic measures are ME:0.04, RMS:1.2, RMSS: 0.8, MS:0.007. Our research indicates the importance of using statistical and cartographic comparison of spatial interpolation models before determining which methods to depend on. The height of stations and the area's elevation must be considered to improve the performance of spatial interpolation methods