Abstract: The BmnRoot software package plays a crucial role in the BM@N NICA experiment and is used both for simulation and event reconstruction purposes. In this article, few approaches to optimization of simulation and event reconstruction algorithms are presented. Results of performance studies of the BmnRoot modules are obtained using software dynamic analysis. Different methods of optimization including parallelization are considered. Algorithms of tracks global matching are discussed. Results of optimization of the virtual planes method are given. Machine learning methods are analyzed from the point of view of efficiency of particle identification. Flexible and optimal extension of the quality assurance module of the BmnRoot is described. Analysis of fragments for SRC studies with the BM@N experiment is also discussed.