Abstract: Existing methods for visualizing underwater objects in three dimensions are usually based on displaying the imaged objects either as unorganised point sets or in the form of edges connecting the points in a trivial way. To allow the researcher to recognise more details and characteristic features of an investigated object, the visualization quality may be improved by transforming the unordered point clouds into higher order structures. There are many algorithms for constructing meshes from point clouds, some of which are more suited to processing data obtained from particular sources. This article presents the application of several methods for generating 3D models from point clouds for the purpose of reconstructing underwater objects, such as shipwrecks. The article presents the results obtained with each method and discusses possible ways of improving the quality of produced meshes.