Predicting the handover success rate in each cell and overall in a cellular network well in advance can help cellular systems to apply timely corrective actions. This can help to ensure high handover success rates for achieving seamless connectivity and quality of service for mobile users. This paper proposes a novel approach to predicting the handover success rate of cellular networks with the help of machine learning techniques employing synthetic data. By leveraging machine learning algorithms and synthetic data generation techniques, datasets generated for the considered networks were utilized for training predictive models. This approach holds the potential to significantly improve the accuracy of handover success rate prediction, consequently leading to enhanced network performance and user experience. Based on the synthetic data generated, the technique also entails prediction of successful calls and dropped calls for a particular cell pair in a cellular network. The model and the approach can be seamlessly applied for real-life operator data for arriving at accurate predictions for analysis and prevention.