An in-depth analysis of the Performance of Random Forest classifiers is presented in this research paper in the context of the classification of real-time power. Accurate classification is essential for effective monitoring and control of power quality disturbances in electrical systems. In this study, a comprehensive real-time power quality dataset was leveraged to assess the efficacy of Random Forest classifiers with regard to various aspects, including feature selection, execution time, and classification accuracy. The impact of different feature subsets, denoted by K, on classifier performance was systematically explored. A corresponding increase in the execution time of the Random Forest classifier was observed as the number of features (K) increased from 20 to 128. However, this trade-off was balanced by a notable enhancement in classification accuracy. An impressive accuracy of 97.50% was achieved when utilizing K = 109 features, requiring 8.8 s of processing time, albeit an accuracy of 97% was obtained with K = 128 in 13.35 s. The effectiveness of power quality disturbance classification tasks being handled by Random Forest classifiers is underscored by this research. A promising choice for real-time power quality monitoring and classification is presented by their exceptional accuracy, even with a moderate number of features, when compared to other commonly used classifiers.