<p>This paper investigates a two-stage hybrid flow shop scheduling problem arising in semiconductor chip coating and etching processes. This problem is characterized by sequence-dependent setup times and various job processing repetitions, to minimize the makespan. A mixed-integer programming model is constructed to jointly optimize job assignment and sequencing. By relaxing certain constraints and structural properties, we propose two lower bounds. Leveraging the Minimal Matrix Setup Time Rule (MMSTR) and an enhanced Kim algorithm, we develope a constructive heuristic for initial job allocation and sequencing. Specialized scheduling rules and a Variable Neighborhood Search (VNS) are employed to refine the solution. Comprehensive computational experiments validate the proposed approach, benchmarking its performance against an exact solver and state-of-the-art algorithms. Results indicate that the heuristic-based VNS achieves an approximate 9% improvement over existing methods and provides solutions within 18% of the best-known benchmarks for large-scale instances. This study provides practical methods for optimizing semiconductor manufacturing operations, improving process efficiency and enhancing resource utilization.</p>