Precipitation is a condensation process on water from the atmosphere that causes rain. Rain has positive and negative impacts on society, especially because of global warming. Estimating precipitation at specific locations is needed to mitigate the negative consequences. Utilizing satellite data is the best way to estimate precipitation, but it has a coarse resolution. Therefore, a study was conducted on spatial downscaling for increasing spatial resolution of precipitation data from the tropical rainfall measuring mission satellite using machine learning in Riau Province, Indonesia. We compared machine learning models namely decision trees, multiple linear regression, support vector machines, and random forest to downscale the data. We used variables like normalized difference vegetation index, digital elevation model, and land cover as the input for the model. Also, we validated the result with the measurement of the rain gauge station at Riau Province Indonesia. Based on the study, we found that the decision tree is the best model to downscale the precipitation data in Riau Province with the mean square error value of 0.00048 and the R2 value of 0.67107. We found that the digital elevation model is the most important variable in downscaling the precipitation data.