The study investigates the relationships between large-scale atmospheric circulation (represented by circulation indices and 27 circulation types derived from gridded mean sea level pressure) and daily precipitation amounts over three regions in the Czech Republic (Central Europe) with different precipitation regimes. We examine how ENSEMBLES regional climate model (RCM) simulations driven by re-analysis reproduce the observed links and capture differences in the links between given regions (lowland vs highlands) and seasons. We study the links of circulation to (1) mean precipitation over the regions, (2) probability of wet days, and (3) probability of extreme daily precipitation (exceeding threshold defined by a high quantile of precipitation distribution in a given season). We find quite strong links in the observed data between atmospheric circulation and these precipitation characteristics. The links are generally more pronounced for highland than lowland regions. More wet days and higher precipitation amounts are found for cyclonic and stronger flows, and for westerly and north-easterly flows. The RCMs are generally able to capture basic features of the links. They nevertheless have difficulties to reproduce some more specific features and differences in the links between the regions. Results also suggest that good performance in some precipitation characteristics may be due to compensating errors rather than model's perfection.