The GaoFen-5 satellite is equipped with a greenhouse gas monitoring instrument (GMI) and a directional polarization camera. Both devices have their own advantages as well as limitations in cloud detection. This study proposes a novel collaborative cloud screening algorithm that uses data from both devices to improve the efficiency of cloud screening for greenhouse gas retrieval. This algorithm is tested with 77581 GMI observation points from the global 16-day on-track measured data, and 9508 clear-sky observation points, i.e. 12.26% points are screened. With the fused moderate resolution imaging spectro-radiometer cloud mask and cirrus reflectance dataset, the validity of cloud detection by the proposed algorithm is confirmed. The accurate rates of cloud detection of 92.93% and 81.91% over land and oceans are obtained, respectively.