The importance of Spironolactone in congestive heart failure (CHF) treatment has been well established. The prediction of the hyperkalemia in the patient using Spironolactone is still not clearly defined. The aim of this study is to develop an accurate prediction model of hyperkalemia incidence in CHF patients on Spironolactone using machine learning techniques. A classification and prediction process have been applied on patients' data of the cardiac center of National Guard Health Affairs, King Abdulaziz Medical City (KAMC). The study was conducted on the records of 1533 patients representing the CHF patient's during the period of 2011-2016. Our experiments show that the JRip classifier achieves the best performance for the Precision (0.983), Recall (0.983), F-measure (0.976) and Accuracy (98.27) metrics while the Naiive Bayesian classifier achieves the best performance for the Specificity (0.652) and AUC (0.93) metrics.