Traffic safety has always been an important issue in sustainable transportation development, and the prediction of traffic accident severity remains a crucial challenging issue in the domain of traffic safety. Based on the severity we can create awareness of happening the accidents according to the zone or place. So the goal is to predict the accident severity. Here we used classification models to find the f1 score of each model and select the best model with a high f1-score to predict the approvals. we take the input parameters, and we train data and the output will be the accident severity.