Cloud data centers frequently face dual challenges situations to hold the stability of labor quantity and maximize using sources because of the dynamic and heterogeneous nature of the applications stored. To reply to those limitations, this observe affords a hybrid VM making plans framework combining the Bayesian clustering with Particle Swarm Optimization (PSO). The first version applies to institution missions in step with the similarity of the workload and the traits of the call for for sources, even as the Bayes opportunity idea is used to refine the host choice and limit overload risk. After that, PSO become used to search for an ultimate making plans answer in a repeated manner through the use of fitness functions to check the time of feedback, electricity performance and use of sources. A matrix -primarily based totally allocation version is brought to symbolize the making plans states and manual the very last implementation selections in every repetition. Test effects display that the approach of presenting development of load stability, decreasing execution delays and enhancing the enlargement of the system, as a consequence making sure performance and flexibility withinside the multi - place cloud environment.