Efficient management of mainstream traffic flow on freeway networks is a critical challenge in urban transportation, with significant implications for congestion mitigation and environmental sustainability. The purpose of this study is to address the problem of predicting traffic volumes and maintaining flow rates below critical densities, thereby preventing the onset of congestion on interconnected freeway systems. Motivated by the need for real-Time traffic control strategies, this research employs machine learning algorithms to forecast traffic volumes, leveraging a comprehensive dataset of traffic patterns on freeways. In our approach, we conducted a comparative analysis of two advanced machine learning algorithms: Long Short-Term Memory (LSTM) networks, which are adept at modeling time-series data with long-range temporal dependencies, and Random Forest regression, known for its robust performance across diverse datasets. We enriched the traffic data through feature engineering, incorporating temporal variables, vehicular counts, and a calculated measure of proximity to critical density for the targeted freeway. Our findings indicate a markedly disparate performance between the algorithms. The LSTM model showed a moderate ability to capture the variance in traffic flow, with an R2 score of 0.619. In contrast, the Random Forest model demonstrated exceptional predictive accuracy, achieving an R2 of 0.998, and substantially outperforming the LSTM model in terms of both Mean Squared Error and Root Mean Squared Error.