Limpets are keystone grazers on rocky shores. We studied spatial patterns of four co-occurring patellids (Patella ulyssiponensis, Patella depressa, Patella rustica, Patella vulgata) and one siphonariid (Siphonaria pectinata) across rockpools and their surrounds. Our approach considered each rockpool as a concentric system of three micro-habitats where limpets were censused (species, counts, and size of individuals): inside rockpool, edge zone, and surrounding open-rock. The whole intertidal zone was sampled on six rocky shores with contrasting landscape variability (amount of open-rock and other micro-habitats) in Southwest Portugal. Additionally, open-rock surfaces far away (>25cm) from rockpools were surveyed on the mid-shore. Three groups of predictors (“physical, habitat-composition, and connectivity variables”) were assessed for each rockpool system. Limpets of various sizes of the three most common species—P. ulyssiponensis, P. depressa, and S. pectinata—were found to coexist inside and outside rockpools on all shores, with rockpools extending their vertical distribution upwards. Patella vulgata was rare; P. rustica was absent from rockpools and their edges. Intra-specific connectivity (correlations in density of the same species between adjacent micro-habitats) showed a decreasing gradient with distance from the pool, being sharper for P. ulyssiponensis (a largely rockpool-resident) than for P. depressa and S. pectinata (suggesting their emergence from rockpools to edges and surrounding open-rock). Abundance and size structure of both P. depressa and S. pectinata were similar on mid-shore open-rock surrounding rockpools and far from rockpools. Landscape variability explained distribution of P. depressa (shores with less open-rock had more and smaller limpets). Habitat-composition (assemblages and substrata) was more important than physical predictors in explaining among-rockpool variability of each limpet species. Densities of P. ulyssiponensis and P. depressa inside rockpools were positively associated with cover of coralline algae and negatively associated with sand cover. Less explanatory power was found in predictive models of S. pectinata, suggesting a more opportunistic nature. There was limited evidence of inter- or intra-specific competition of the three species in rockpools and open-rock. Competition between P. ulyssiponensis and P. depressa was most likely in rockpool edges. Rockpool-generated mosaic landscapes are linked by intra-specific connectivity of limpet populations and modulated by interactions among different limpet species.