Proposes a method for target tracking by underwater robots. How the dynamic properties of the target objects can be used to navigate autonomous underwater vehicles (AUVs) is studied. The objects of interest are extracted by using optical features, and their dynamical properties are calculated using optical flow techniques. Using past and present information, the consecutive dynamic behavior of the object is estimated. The paper presents a sensor fusion scheme to derive the navigational commands for the AUV. The results of the extraction of the target in the image, based on optical flow, and estimation of the object speed are presented with experimental results. These results demonstrate that the performance of the proposed target tracking algorithm is a useful tool for underwater robots.