Rare earth is a vital strategic resource whose price fluctuations exert a substatical influence on the national and corporate economy, so predicting the price of rare earth in advance is also important for the adjustment of industrial scale. Currently, the forecasting of rare earth element prices predominantly relies on time series analysis. This study, however, diverges from the conventional approach by employing an areay of advanced machine learning and neural network algorithms. Such as Regression analysis, Decision tree, XGBoost, LSTM to predict rare earth price respectively. From the perspective of R-squared and Mean Absolute Error (MAE), the effect is better than that of a single time series pre-diction method, which can be widely applied to the price prediction of a series of resources such as rare earths.