A number of imaging methods use phase information from received signals to construct an image. Each method suffers from a phase ambiguity problem; that is, it is impossible to distinguish between phases that differ by 2p. A number of algorithms have been proposed to use information from surrounding data points to determine the actual phase from the “wrapped” phase.In this paper, we propose a new method of two-dimensional phase unwrapping based on supervised neural network techniques. This technique is shown to be able to detect the correct unwrapping locations where other conventional techniques fail. The results of this new phase unwrapping algorithm are shown using optical Doppler tomography velocity images.