The phase recovery and dephasing of single-frame structured light images are crucial in the field of three-dimensional (3D) imaging. Theoretically, short-time Fourier transform (STFT) can efficiently complete phase recovery under the condition of selecting appropriate parameters; however, the prerequisite of "appropriate parameters" is often challenging to achieve due to the lack of prior information. Herein, a single-frame 3D imaging algorithm of structured light field based on adaptive STFT is proposed using a structured light field system to take full advantage of the characteristics of light field imaging. The suggested technique fuses the structured light field image from three angles of texture, instantaneous frequency, and depth estimation information and further divides the light field image into several regions that meet the STFT constraint. The associated STFT parameters are estimated for each zone using the instantaneous frequency, and then the relative phase is recovered. Next, the phase separation method of the structured light field is used to achieve 3D imaging. The experimental results show that the proposed adaptive STFT algorithm can successfully accomplish the 3D imaging of the measured object.