The VN-Index in the Vietnamese stock market not only reflects the health of the national economy but is also an important channel for capital mobilization. VN-Index accurately forecasting is very important for investors, policymakers and analysts. This is because the index is strongly influenced by many complex factors, intertwined between the domestic and international context, from economic policies to market psychology and global trends. This study focuses on clarifying the multidimensional factors affecting the VN-Index. The analysis focused on the effects of global stock market performance, exemplified by the S&P 500, alongside the influence of geopolitical risks and changes in gold and oil prices. Data spanning from January 2002 to April 2024 were employed to examine the difficulties encountered when predicting the VN-Index. Two linear regression techniques, Ordinary Least Squares (OLS) and Lasso, were utilized in this assessment. The Lasso method, which is notable for its ability to handle multicollinearity and minimize overfitting by optimization, helps improve forecasting accuracy by reducing the weight of less important variables. The analysis process also includes testing different lambda parameters, applying sliding window analysis, and using evaluation indices such as MAE and RMSE to tune the model and enhance the reliability of the prediction. The results of the study provide deeper insights into the determinants of VN-Index volatility and demonstrate the effectiveness of advanced econometric methods in modeling financial markets.