Species distribution model (SDM) has conventionally been used for evaluating the distribution of single species, but comparisons between different SDMs are possible for evaluating the geographic similarity between taxa. Here we used a parasite and host system to infer the geographic overlaps between species with tight biological interaction, e.g. parasites and their obligate host; specifically, we used the horsehair worm Chordodes formosanus and its three different mantis hosts to study the extent of niche overlap. We retrieved presence points for the host species and the parasite and we built the SDMs with MaxEnt implemented in ENMeval by using selected bioclim variables (based on VIF values) at a 30 seconds scale. The models showed that the hosts and parasite do not occur in the high elevation areas in Taiwan, which was expected based on their biology. Interestingly, the predicted parasite distribution included areas without collection records, implying local extinction or sampling bias. We subsequently evaluated niche overlap between hosts and the parasite according to five similarity indices (Schoener’s D, I statistic, relative rank, Pearson correlation coefficient and the rank correlation coefficient rho). Our models showed high similarity of SDM predictions between hosts and the parasite. There were differences among metrics about which host shared the highest similarity with the parasite, but the majority of the results indicated that the Japanese boxing mantis has the highest niche similarity with the horsehair worm. The choice of the niche overlap metric to use can be seen as a way to get informations on the parasite’s ecology, which can be important for endangered species SDMs are reliable tools for host and parasite conservation management and could help to improve biological and ecological knowledge of parasites.