One of the applications of machine learning in predicting medical insurance prices considering health and economic factors is because this branch analyzes how healthcare resources are allocated and how healthcare outcomes are determined. The production of medical insurance prices encounters challenges rooted in data accuracy and ethical consideration of machine learning models. In this paper, we proposed an efficient framework for predicting medical insurance prices and a delicate balance between accuracy and fairness to ensure the efficacy and ethical soundness of the pricing process using five machine learning algorithms MAPE , R2. On four different datasets, Cross-validation number of folder:5 and the best result on MAPE is a tree with the smallest number of errors was 3.9%, Cross-validation number of folder:10 and the best result on MAPE is a tree with the smallest number of errors was 3.5%, Random sampling training set size 80% and testing 20% the best result on MAPE is a tree with the smallest number of errors was 4.1%,%, Random sampling training set size 90% and testing 10% the best result on MAPE is Tree with the smallest number of errors was 4%. The best result of all datasets on MAPE is Tree.