This study evaluated the impact of different integration methods when using hybrid models for time series forecasting, employing monthly global oil price data from January 2004 to December 2023. The research included individual models of Autoregressive Integrated Moving Average (ARIMA) and Support Vector Regression (SVR) and a hybrid model (ARIMA-SVR) utilizing multiple integration techniques (additive, multiplicative, and regression). The results demonstrated the superiority of the additively integrated hybrid model, which achieved the lowest values for forecast accuracy metrics (MAE, MPE, MAPE, and MSE), significantly outperforming the other models. Specifically, this model showed a 46.4% improvement in MAE compared to the ARIMA model and a 29% improvement compared to the SVR model. The regression hybrid model followed in performance, followed by the multiplicatively integrated model, the SVR, and lastly, the ARIMA model. These findings highlight the effectiveness of hybrid models, particularly those with additive integration, in enhancing the forecasting accuracy of complex time series exhibiting both linear and nonlinear patterns. The study recommends exploring more sophisticated integration methods and expanding the scope of applications in future research.