We developed a global dataset of downscaled future projections developed by applying a statistical method for climate model downscaling and bias correction. To develop the dataset, we applied the delta method, which comprises the sum of interpolated anomalies of each GCM to the WorldClim 1-km spatial resolution dataset. The GCMs were the 35 Coupled Model Intercomparison Project Phase 5 (CMIP5) models, for four representative concentrations pathways (RCPs). For each of these, we used the 30-year future periods named as 2030s (mean of 2020-2049), 2050s (2040-2069), 2070s (2060-2089) and 2080s (2070-2099) with three climate variables (mean monthly maximum and minimum temperatures and monthly rainfall). From these, we also derive a set of bioclimatic indices.