This paper introduces a novel method of successful processing the large three-dimensional (3D) spatial data for Virtual museum design. Billion triangular faces are obtained from the 3D scanning process of national clothes. In order to avoid the filling holes problem we have included additional steps in post scanning process. Removing isolated vertices and all redundant faces, in the first step algorithm solves all topological error issues. This step also includes mesh optimization for two purposes: i) removing duplicated edges and faces and ii) mesh simplification for a precise unwrapped map. Algorithm in the next step performs texture creation and UVW mapping. Separated fields of the mesh faces contain rich and unique textures analogously to original national clothes. Separating these fields by color, texture and curvature features, we also achieved better rendering performance in the 3D web preview. This step is the final step in the presented algorithm. Classifying texture pixels in database in strict correlation to the real museum cloth and jewelry classifications, we have designed powerful software engine. Algorithm is crucial for the Virtual museum in order to allow the simple and efficient use and administration through user view and administrator panel.