Home insurance provides essential financial protection against potential damages to insured properties. While customers initially choose specific coverages when purchasing a policy, insurance companies have noted that many clients face the decision of whether to renew their contracts or let them lapse. During the renewal process, some customers modify their coverage by adding protections, while others may choose not to renew due to factors like premium costs or changing needs. Understanding the factors that drive these renewal decisions is crucial for insurers to improve customer retention strategies. This study aims to analyze the factors influencing customers’ decisions to renew or not renew their insurance contracts. By examining the initial coverages chosen at the contract’s inception and any modifications made during the renewal process, the research seeks to identify key determinants that drive these renewal decisions. With the growing role of artificial intelligence (AI), particularly machine learning (ML) and data analysis, this research aims to leverage these technologies to better assess risk factors and predict customer behavior. By employing machine learning models to analyze large datasets, insurers can enhance their ability to encourage clients to opt for more comprehensive coverage from the outset, thereby remaining competitive in a dynamic global market. The objective of this study was to predict whether customers would cancel or renew their home insurance policies. This decision often arises when a policy expires, typically due to reasons such as non-payment or failure to renew within the contract’s timeframe. Using a real-world dataset from home insurance companies, spanning the years 2007 to 2012, key policy attributes, building characteristics, geographical zones, coverage benefits, and risk factors were analyzed. Various machine learning algorithms, including Decision Tree (DT), K-Nearest Neighbors (KNN), Support Vector Machine (SVM), and XGBoost (XG), were applied to the dataset, which was divided into 70% for training and 30% for testing.