This research presents a dynamic scheduling model designed to reduce extreme departure time deviations for freight train requests and distribute deviations more evenly across the planning horizon, as an improvement over the current first-come, first-served method. By integrating deterministic demand forecasting and congestion-based penalties into a dynamic scheduling algorithm, the model evaluates overall customer satisfaction after allocating all requests. A sensitivity analysis is conducted to identify the optimal balance between minimizing forecasted congestion and reducing current departure time deviations. The approach aims to achieve a scheduling strategy that minimizes extreme departure time deviations without disproportionately impacting overall time deviations. Results demonstrate the model's ability to reduce some extreme deviations in low-capacity network scenarios and distribute the deviations more evenly throughout the planning horizon, with minimal impact on overall time deviations. In high-capacity networks, improvements are negligible. The findings highlight the potential for targeted benefits under constrained conditions while underscoring the need for further refinement to ensure consistent and meaningful performance across diverse scenarios.