Visualization techniques have been greatly developed in the past few years. Three dimensional models based on satellite and aerial imagery are now being replaced by models generated on aerial laser scanning (ALS). The most modern of such scanning systems have the ability to acquire over 50 points per m2 and to register a multiple echo, which allows reconstruction of the terrain together with the terrain cover. However, ALS data accuracy is greater than 10cm and the data is often incomplete: there is no information about the ground level (true for most scanning systems)and often about the facade or objects which are covered by other objects. There is also no data on beneath the object. However, the terrestrial Laser Scanning (TSL) obtains higher accuracy data (1÷5cm) on all of those elements which are incomplete or not visible with ALS methods (facades, complicated structures, interiors, bridges, etc.). This paper presents chosen aspects and methods for combining data from aerial and terrestrial laser scanning for the purpose of creating three-dimensional models of cities. This will be done based on data acquired using the Lite Mapper Q680i aerial scanner and the ScanStation2 terrestrial laser scanner. Methods for choosing tie points to combine point clouds in different data (UTM from aerial measurements and a local datum from terrestrial scanning) will be analyzed. The results of transformations will also be evaluated. The problem of modeling point clouds and combining created models with othro-images will also be dealt with.