This study aims to develop a machine learning-based sales forecasting model to reduce the uncertainties that companies face in their inventory and production planning by analyzing the impact of special days on retail sales. The sudden increases in demand, especially during periods such as New Year's, Black Friday, and Christmas, can create operational risks for businesses if accurate forecasting models are not used. In this context, a forecasting model was created using the Random Forest algorithm with Amazon's OLED television sales data, and it was demonstrated how this model can be used to understand sales behavior on special days. The findings of the study reveal that machine learning-based models can be effectively used in sales forecasting processes and contribute to the data-driven decision-making processes of businesses.