This study introduces a hybrid forecasting framework that integrates Vector AutoRegressive Integrated Moving Average (VARIMA) with machine learning models to improve multivariate time series prediction in the electronics sector. The dataset comprises real-time sales records enriched with multiple explanatory variables, including Year, Month, Promotional Offers, Weather, Pricing, Sales Policies, and Eid/Festival impacts, all of which strongly influence consumer demand and introduce pronounced seasonality. The VARIMA component is employed to capture linear dependencies and seasonal trends, while machine learning algorithms are applied to the residuals to model nonlinear relationships and hidden interactions among variables. By combining the strengths of both statistical and data-driven approaches, the proposed hybrid method delivers superior forecasting accuracy compared to standalone models. The findings provide actionable insights for demand planning, promotional strategy, and policy adjustments, enabling firms to better align supply chain decisions with dynamic market conditions.