Cloud overlapping seems to be a frequently occurring phenomenon as surface-based radar shows. Since most of the satellite remote sensing techniques of clouds assume a 1-layer cloud system for the retrieval, to detect overlapping cloud cases accurately is very important for improving the quality of the products. It is, however, quite difficult to detect it from space-borne passive radiometer. Several algorithms that use infrared (3.7, 11 and 12 um, and 15 um CO2 bands) and near-infrared and infrared (1.6 and 11um) have been developed so far with mixed operational success. Here we demonstrate another method using visible (0.64um), solar-infrared(3.7um), infrared (11um) and split-window (12um) wavelengths. We assume a 2-layer cloud system comprising thin cirrus overlying a thick layer cloud in this study. The present algorithm first applies the VIST (Visible, Solar infrared and Infrared Technique) to satellite imager data. The VIST retrieves the cloud optical depth t, effective particle radius re and cloud temperature Tc for each pixel assuming 1-layer cloud system. We then compute the brightness temperature difference between 11 and 12um (BTD) for the retrieved t, re and Tc , then compare the computed BTD with observed BTD. These BTD values should often be significantly different in case of the assumed 2-layer system. We present some numerical simulations and analyze GOES data. The results are validated with 35GHz radar data taken at the ARM/SGP site.