A real-time computer-generated moiré profilometry (CGMP) with adaptive filtering algorithm is proposed. In CGMP proposed previously, only one static sinusoidal grating needs to be projected onto the measured object and only one deformed pattern caused by the object needs to be captured, which is ideal for real-time three-dimensional (3-D) measurement. By means of generating moiré fringes between the captured deformed pattern and the prestored fringe patterns on the reference plane via computer, the 3-D shape of the measured object can be reconstructed successfully. In phase calculating, this method needs to extract the DC component by filtering. If filtering operation required manual intervention each time, its real-time measuring feature may be limited. For this, an adaptive filtering algorithm based on the maximum fundamental frequency component proportion evaluation function is proposed. It can automatically retrieve an object’s phase information according to the Fourier spectrum distributions of different measured objects without any manual intervention. Experimental results demonstrate the feasibility and validity of the real-time 3-D measuring application.