Churning the customer has become an emergent issue in the telecom world due to its direct influence on revenue and profitability. Prediction of churn allows telecom firms the ability to act ahead to reduce customer loss. Thus, in this research we have used machine learning algorithm-Decision Trees and Random Forests-to predict churn data. These models are able to handle large datasets and grasp complex patterns in the behavioral actions of customers. This research applies a dataset with 7,044 customer records and 21 features: demographic information, usage patterns, and churn status. Preprocessing of data was carried out and two Decision Trees and Random Forest models were trained to find out key factors affecting churn in customers. Both models were assessed on accuracy, Precision, Recall, and F1-Score metrics. An accuracy score provides an overall measurement. Precision and Recall are specific performance measures that are useful as the churn dataset is biased. The F1 Score is a harmonic mean between Precision and Recall, taking into account the tradeoff with false positives and false negatives. Our results show that both models are efficient for the classification of churn; however, Decision Trees have performed better than Random Forests. Decision Trees have proved to be effective since they prevent overfitting and handle feature variability in such a way that increases prediction accuracy. This helps telecom companies identify at-risk customers so that interventions can be implemented on time and targeted retention strategies can be put in place to enhance customer satisfaction and lower churn rates. customers.