Maintaining track continuity and correct target identities is a key challenge in maritime radar tracking, particularly during atypical scenarios such as close-proximity encounters, erratic vessel movements, or temporary occlusions. Generalized tracking algorithms, commonly used in conjunction with coastal radar surveillance, often fail under these conditions, resulting in fragmented tracks and identity swaps. This study presents a modular, scenario-specific optimization framework for radar tracking, designed to optimize performance of trackers to a specific scenario. The study utilizes the VIMMJIPDA algorithm as the core tracker while applying automated hyperparameter tuning. The framework leverages a custom performance metric that integrates positional accuracy and track continuity, enabling automated optimization of hyperparameters tailored to each scenario. Evaluation was conducted using a GPS-validated stress-test dataset featuring two vessels performing coordinated, high-maneuverability maneuvers. The optimized tracker consistently preserved track identity and continuity, including the smaller, more agile vessel, while maintaining positional errors within operationally acceptable limits. Linear mixed-effects modeling revealed a statistically significant but practically modest increase in point-wise positional error relative to the baseline. These findings demonstrate that scenario-specific optimization of radar trackers can effectively mitigate track and identity loss under challenging scenarios while preserving operational accuracy.