The Quran is considered the holy book for the Muslim community, and its text contains vast amounts of knowledge and guidance that Muslims strive to extract and understand. To achieve this, Quranic text classification plays a crucial role in categorizing and organizing the vast amount of information contained within the Quranic text. However, the process of Quranic text classification is not without its challenges. One of the significant challenges in Quranic text classification is obtaining a homogenous and balanced dataset to train the classification models accurately. Due to the nature of the Quranic text, which contains various topics and themes, the distribution of Quranic text classes is often abnormal. This abnormal distribution makes it difficult to obtain a consistent and balanced dataset, which can weaken the overall classification performance. To address this issue, this paper proposes a new oversampling method that employs Genetic algorithm to generate an optimal and balanced dataset simultaneously. The proposed method is specifically tested for Quranic topics that contain several imbalanced binary classes. The results of the study demonstrate the effectiveness of the Genetic algorithm in generating a balanced dataset, which leads to better classification performance results.Overall, this paper highlights the importance of Quranic text classification and the challenges associated with it. The proposed oversampling method provides a novel solution to address the issue of imbalanced Quranic datasets and can significantly improve the accuracy of Quranic text classification.