Tourism plays a vital role in global economic development by contributing significantly to employment generation and gross domestic product (GDP). Tourism data are collected in large amounts and are subject to seasonal effects and complex patterns. Hence, precise forecasting is quite challenging. In this context, this study proposes a predictive model that mines historical tourism data to uncover key determinants for tourism sales. The key technique is based on the Random Forest regression algorithm, which is widely used for predictive tasks due to its superior accuracy and reliability. Moreover, this technique is highly capable of capturing non-linear relationships among various determinants such as hotel prices, transportation costs, and historical sales records. The proposed model follows a series of processes to enhance the quality of predictions based on data mining techniques. The proposed model is evaluated based on various metrics, including Mean Absolute Error (MAE) and R-squared score, which is widely used for regression-based predictive models. The proposed model is found to yield superior accuracy and is highly beneficial for decision-makers in the tourism industry. This research also focuses on the development of a user-friendly predictive system, where in the users can input the required parameters and obtain real-time sales forecasting results. The usability of the system can be improved by adding visualization tools like trend graphs and performance indicators, which will help the users understand the results of the forecasting process better.