The CAMS radiation service (CRS) provides information on surface solar irradiance (SSI). The service implements physical models McClear and McCloud with a fast parametrization of radiative transfer in the atmosphere as part of the Heliosat-4 SSI retrieval scheme (Qu et al., 2017). The current CRS v4.5 operational processor uses cloud properties derived from images of Meteosat Second Generation (MSG) satellite instruments in their 15-minutes resolution using APOLLO_NG (Klüser et al., 2015). In APOLLO_NG, the cloud masking is based on Bayesian based probabilistic approach and for the cloud optical thickness retrieval, the 0.6 micrometer channel is used. CRS is subject to a continuous development and the recent evolutions have focused on the application of the APOLLO_NG cloud retrieval for the calculation of MSG SEVIRI based irradiances (Schroedter-Homscheidt et al., 2022) and for the instruments aboard third generation geostationary satellites HIMAWARI-8/9 and GOES-16, in preparation for MTG. HIMAWARI-8/9 is currently being integrated in the CRS operational processing chain as a first step towards the extension of the service to a global scale. A regular monitoring and evaluation of all input and output parameters is part of the CRS activities and ensures the service quality. Clouds are the main cause of SSI variability and also the dominant source of errors in the all sky SSI retrieval. The most relevant cloud properties for the SSI retrieval are cloud mask, cloud optical thickness (COT) and cloud phase. While cloud products have been widely evaluated in the cloud retrieval community, an overall assessment for all retrieval conditions and splitting into favorable and less favorable retrieval conditions is not very well documented, but required by solar energy users. This study will present the evaluation of the APOLLO-NG cloud properties through comparison with the CLAAS-3 (CLoud property dAtAset using SEVIRI - Edition 3) CPP (Cloud Physical Properties) database from CMSAF (Satellite Application Facility for Climate Monitoring) services, for different locations in the MSG field of view. As Heliosat-4 offers the possibility to use cloud properties from different sources, the analysis will include evaluation of CRS SSI obtained from both APOLLO_NG and CPP versus ground observations of SSI. Such an analysis will help assess the suitability of existing cloud retrievals APOLLO-NG and CPP for the use in surface irradiance retrieval schemes and for all retrieval conditions as required by solar energy users.