Traffic accidents have grown rapidly throughout the world in recent years, causing great loss of life and property. Therefore, predicting traffic accidents is very important for improving transportation and public safety. Machine learning (ML) is a subfield of artificial intelligence that can extract information from dataset and use statistical approaches to predict values. In this study, three ML techniques were applied to predict the number of dead or injured in traffic accidents in Turkey until 2029: linear regression, decision trees, and random forest. These techniques were tested using a real dataset obtained from the TUIK website. In the results, it was seen that linear regression (LR) had the best performance. This result shows the superiority of the approach in predicting road accidents. Ultimately, this study will help road transport and insurance agencies develop road safety strategies.