This paper tackles the computational hurdles in region-based Petri net (PN) synthesis, focusing on the time-consuming identification of minimal regions essential for accurate model construction. Given the limitations of existing algorithms in handling complex systems, we introduce an improved algorithm that integrates advanced multiset expansion techniques to enhance the generation of minimal regions significantly. This approach accelerates the computation process while maintaining the accuracy and robustness necessary for effective PN construction [5]. We contextualize our contributions by reviewing recent advancements in region-based discovery algorithms, highlighting our algorithm’s improved capability to construct k-bounded PN with complex behaviors such as concurrency prevalent in process mining. Comparative analysis and empirical validation show that our algorithm outperforms existing methods in speed and scalability without sacrificing detail in process dynamics representation. This research contributes to both the theoretical framework of PN synthesis and the practical aspects of process mining tool design.