Although turtles are the most threatened taxonomic group within the reptile class, we have a very limited understanding of how turtles respond to climate change. Here, we evaluated the effects of climate changes on the geographical distribution of Caspian pond turtle (Mauremys caspica). We used an ensemble approach by combining six species distribution models including artificial neural network, generalized boosted model, generalized linear model, flexible discriminant analysis, random forest and multivariate adaptive regression splines. To predict the future distribution, modelling projection of MRI-CGCM3 was used for the year 2070 under four scenarios of representative concentration pathways (RCP). Based on the findings, the suitable habitat of Caspian pond turtle was estimated to be about 835941 km2 (about 8.73%) of the study area. Our model projections showed that about 26 to 33% of the current suitable habitats will be unsuitable by 2070 due to climate change. The annual precipitation (24.56%), precipitation of wettest quarter (24.28%), precipitation seasonality (16.93%) and temperature seasonality (14.64%) had the highest contribution to model performance of Caspian pond turtles. Overall, our findings emphasize the need for a comprehensive understanding of the complex effects of climate change on the species, specially turtles.