This paper centers on the problem of predicting the volatility of financial markets and assessing their risks. Firstly, a financial market forecasting method based on EEMD-LSTM is realized. The realized volatility forecasting model based on EEMD-LSTM is constructed with the realized volatility series of CSI 300 index as the research object. And RMSPE, R² and MAE are used as the evaluation indexes of the model’s forecasting effect to compare the experiments with the comparison model. Then the VaR calculation method based on Copula function to improve the traditional Monte Carlo simulation (MC) method is proposed for financial market risk assessment. Applied to the trading data of PetroChina stock market, the calculation results show that the Copula-MC method is more accurate and effective in approaching the actual VaR.When the EEMD-LSTM-based financial market prediction method and the risk assessment method based on Copula-MC are used in combination, the effect of estimating VaR is more satisfactory, and the method in this paper can effectively predict the volatility of the financial market and compute the value of its risk.
