This paper introduces a octree search Kinetic Monte Carlo(OS-KMC) that combines the topological requirements for representing an evolving multivalued surface using minimal memory and an efficient search algorithm for the realization of the KMC time evolution. In addition, the data structure for OS-KMC also provides a nature way to generate hexahedral element meshes for the integration between simulator and performance analysis tool. The density of mesh grid and the refinement of the hexahedral element can be controlled in this octree based mesh generation method. In simulating surface morphology during wet etching and micro structure formed by composite MEMS processes, the octree search KMC shows good simulation results with better calculation performance. The octree structure enables the Monte Carlo solutions for large scale problems with complex dynamic surfaces.