The invention discloses an adaptive 3D video coding and decoding method based on compressed sensing. The method comprises the following steps: firstly, dividing a multi-view-point video into a key view point and a non-key view point, and dividing the key view point into a key frame and a non-key frame; adopting a higher sampling rate for the key frame; performing image block classification (SKIP, INTER and COMBINED modes) on the non-key frame according to the residual error between the current frame and the key frame, and adopting different sampling rates for different modes; and adopting similar methods for the non-key view point. The operation method for the decoding end comprises the following steps: firstly, restoring the key frame of the key view point; and then restoring the non-key frame by adopting information of the key frame; and restoring the non-key view point by using information of the key view point. According to the method disclosed by the invention, the computation complexity of the coding end can be effectively reduced, the scenes with weak computing power at the coding end and powerful computing power at the decoding end can be adapted, and the compression efficiency of 3D videos can be significantly improved.