Abstract: Multibeam sonars are widely used in applications like high resolution bathymetry measurements or underwater object imaging. One of the significant problems in multibeam sensing of the marine environment is large amount of data which must be transmitted from the sonar processing unit to an operator station using a limited bit rate channel. For instance, such a situation would be in the case when the multibeam sonar was mounted on the autonomous underwater vehicle operating on large depths and transmitting the data to the operator station using acoustic channel. In this context, the authors propose a method for multibeam sonar data size reduction. It relies on the use of wavelet decomposition technique combined with run-length and Huffman coding. The method was applied for lossy compression of raw bathymetry data which had been generated by a multibeam sonar processing unit in a form of a set of points in three-dimensional space. The performed tests revealed that without introducing the substantial distortion into the processed bathymetry data, the proposed approach allows to obtain better compression ratios than in the case of using standard lossy JPEG-like compression techniques.