The S transform is one of the time-frequency analysis techniques which combines the advantages of both the short-time Fourier transform and wavelet transform. The one-dimensional S transform has been successfully used in fringe phase demodulation based on structured light projection. The two-dimensional (2D) S transform is able to carry out the time-frequency analysis for images in two directions, which is better than one-dimensional (1D) S transform. The 2D S transform based on the structured light projection method is introduced to the three-dimensional (3D) optical measurement for completing S transform fringe phase demodulation theory. This work studies the theory and application of the 2D S transform in fringe phase demodulation and gives a detailed theoretical analysis. Furthermore, a comparison between the 2D S transform and the 1D S transform is carried out in fringe phase demodulation. The comparison result shows that the 2D S transform will achieve a higher accuracy than 1D S transform in extracting the phase distribution of the fringe patterns, even when the fringe patterns are seriously polluted by noises. Both the computer simulations and experiments verify that the 2D S transform outperforms the 1D S transform in the fringe analysis.