This study evaluates the reliability of merged satellite precipitation products based on single machine learning (SML) and double machine learning (DML) algorithms to estimate meteorological drought events over Kenya from 2000 to 2019. This study selected four SML algorithms, including Random Forest (RF), Support Vector Machine (SVM), K-nearest neighbors (KNN), and the gradient boosting machine (GBM) algorithm, to merge satellite precipitation products. In contrast, the DML algorithm is developed based on the classification RF approach and combined with these regression machine learning models (i.e., RF-RF, RF-SVM, RF-KNN, and RF-GBM). The four gridded precipitation products, including Climate Hazards Group Infra-Red Precipitation with Station (CHIRPS), Integrated Multi-Satellite Retrievals for GPM (IMERG-v06), Precipitation Estimation from Remotely Sensed Information using Artificial Neural Network-Climate Data Record (PERSIANN-CDR) and ERA-5 reanalysis datasets utilized in merging approaches. In total, we compared twelve precipitation products, including four gridded precipitation products (GPPs), four SML, and four DML, against Climate Research Unit (CRU) based rain-gauge (RG) observation. The Standardized precipitation indices (SPI) were estimated from these twelve precipitation datasets and compared their performance of CRU observation across Kenya. In addition, a total of seven statistical metrics were utilized for the performance assessment of all precipitation estimates. The results revealed that the DML merged products achieved higher accuracy (CC = 0.75–0.9) compared to SML products or other GPPs products against observational data. Also, DML merged products show higher performance (CC = 0.8–0.86) for two rainy seasons (long rain and short rain) than other GPPs, and SML merged products, except RF. Moreover, our findings indicate that the DML merged precipitation products identified major meteorological drought events with the lowest RMSE error than other precipitation products based on multiple SPI timescales (SPI-3, SPI-6, and SPI-12). The DML algorithm significantly improves SML models for precipitation estimates to monitor meteorological drought events over Kenya. [ABSTRACT FROM AUTHOR]