To achieve sustainable development of hydroelectric resources, it is necessary to understand their availability, variability, and the expected impacts of climate change. Current research has mainly focused on estimating hydropower potential or determining the optimal locations for hydropower projects without considering the variability and historical trends of the resources. Herein, the hydropower potential variability from reconstructed streamflow series estimated with a non-parametric gap-filling method and geographic information systems (GIS) techniques are analyzed. The relationships between hydropower and large-scale climate variability, expressed by sea surface temperature, are explored. Finally, we project hydropower potential through 2050 using 15 global circulation models with representative concentration pathway (RCP) 4.5. We used four watersheds in central Chile as a case study. The results show significant interannual and inter-basin hydropower potential variability, with decreasing trends over time modulated by alternating positive and negative decadal trends; these modulations exhibit greater intensities than the general trends and are attributable to climatic oscillations such as El Niño. Future scenarios indicate high hydropower availability and a possible over-investment in hydroelectric plants in two of the four studied watersheds. Results show the need to improve the current policies that promote hydropower development including hydropower resource variability in order to achieve optimal, sustainable hydropower development worldwide.