Operating Room Scheduling (ORS) in public hospitals presents a critical operational challenge, where reliance on manual, heuristic-based processes leads to significant inefficiencies, long patient waiting times, and inequitable access to care. This study proposes and validates a novel optimization framework using Constraint Programming (CP) to address this complex problem. We introduce a methodology based on two sequential CP models designed to handle the dynamic and heterogeneous nature of patient arrivals. Using real-world data from a 2024 public hospital network, the models systematically assign various patient types—including scheduled outpatients, newly admitted inpatients, and high-priority cases with legal mandates—to available operating blocks, respecting a rich set of clinical, operational, and hierarchical constraints. The results demonstrate the framework’s viability for daily planning, generating high-quality, feasible schedules in consistently under one minute. A comparative analysis against the formalized current practice (a PRIORITY-FCFS heuristic) for a high-demand instance reveals the CP approach’s superiority: it increases operating room utilization by 6.4%, schedules 7.5% more patients, and drastically reduces the maximum individual waiting time by 40.3%. While the approach faces scalability challenges for weekly planning horizons, it proves highly effective for reactive, day-to-day scheduling. The study concludes that CP provides a robust and adaptable methodology capable of significantly enhancing both the efficiency and equity of surgical scheduling, offering a powerful decision-support tool for transforming hospital operations.