Computer systems must respond to frequently changing user needs in order to remain operational. Their increasing size and operational complexities intend to make them difficult to maintain. A change in a business process is a complicated task especially during the process execution, where a small change can significantly affect the rest of the system with undesirable impacts. In this work, we focus on studying the problem of change impact propagation in Business Process Management (BPM). We propose in this paper an approach that can predict the level of change (LC) in business models. There are three level of changes (Low, Medium, and High) based on structural metrics as used in the predictive model. Five different machine learning (ML) algorithms are used in this model to show their comparison analysis. This issue is important as the LC in business process before implementing any changes helps the organization to make decision in prior. To validate the purposed approach, the experiments conducted in this study show the improved performance using the SVM and Guassian Na ̈ıve Bayes algorithms.