Abstract: For the last decade multibeam sonars have been increasingly used for mapping and visualization of the bottom surfaces to provide the 'physical bases' for environmental studies due to theirs unprecedentedly high resolution mapping ability. However, the raw sonar records are subject to systematic errors, random noise and outliers. In this paper Kalman filtering approach to generating optimal estimates of bottom surface from a noisy raw sonar records is presented. The experiment on the surface indicates that after applying the Kalman filtering technique the outliers of raw records can be efficiently detected and removed. Moreover in the same time, the two-step Kalman filtering method is applied, which aims to filter every multibeam sonar swath and enable 3D seabed visualization in real time. The 3D bottom relief before, and after the filtering method application is also presented.