In this project we deal with surface reconstruction from unorganized data. The cluster creation method for surface reconstruction from a cloud of points is applied as a tool for searching for patterns and organization – data regularity and for the points cloud partitioning. The cluster creation algorithm is based on calculation of dimensions and local densities for each point in the cloud of unorganized points. The obtained triangular mesh is often too detailed. The details are insignificant and often invisible when the surface is rendered. Consequently, the mesh is reasonably simplified.