Context Invasive feral cats strongly threaten native fauna on islands worldwide. The impact of feral cat populations is assumed to be higher in the presence of introduced rodents and may also vary according to an island’s rodent species assemblage. Aims This study assessed feral cat impacts on island biodiversity in French Polynesian islands harbouring differing assemblages of rat species, by investigating their abundance and trophic ecology. Methods We focused on the following six human-inhabited islands of three archipelagos in French Polynesia (South Pacific) with differing rat species assemblages: Tahiti and Moorea, Ua Huka and Tahuata, Rurutu and Rimatara. We studied (1) cat abundance, by setting up camera traps during 15 consecutive days to provide an abundance index, the general index (GI) or mean number of virtual captures per camera per occasion, and (2) cat diet, by performing macroscopic analyses of scat samples to determine the frequency of occurrence (FO) of prey categories. Key results Our study showed previously unreported patterns of cat-abundance index for the Society archipelago, with a GI of 0.30 ± 0.12 in Tahiti and 0.02 ± 0.02 in Moorea; for the Marquesas, with a GI of 0.75 ± 0.20 in Ua Huka and 0.20 ± 0.06 in Tahuata; and for the Austral, with a GI of 0.06 ± 0.04 in Rurutu and 0.19 ± 0.05 in Rimatara. Feral cats are shown to prey strongly on introduced rodents (FO = 91.3%), arthropods (37.8%), squamates (18.6%) and birds (13.3%) in our study sites. FO of birds are particularly high in Rimatara (31.9%) and Ua Huka (16.8%). Conclusions We demonstrated that feral cats represent a serious threat to biodiversity in French Polynesian islands, with 15 species preyed on, including eight endemic birds, four of them being considered threatened by the IUCN Red List. Surprisingly, relative abundances of feral cats were higher both on islands harbouring only one rat species (Ua Huka, with only Rattus exulans) and on islands free of black rats. This finding raises questions regarding the ‘hyperpredation’ hypothesis for multi-invaded island ecosystems. Implications This study on islands with differing assemblages of introduced rodents demonstrated the need for invasive predator studies in multi-invaded ecosystems, so as to improve bird conservation and guide management strategies and site prioritisation.