The effects of climate change on infectious diseases are a topic of considerable interest and discussion. We studied West Nile virus (WNV) in New York (NY) and Connecticut (CT) using a Weather Research and Forecasting (WRF) model climate change scenario, which allows us to examine the effects of climate change and variability on WNV risk at county level. We chose WNV because it is well studied, has caused over 50,000 reported human cases, and over 2200 deaths in the United States. The ecological impacts have been substantial (e.g., millions of avian deaths), and economic impacts include livestock deaths, morbidity, and healthcare-related expenses. We trained two Random Forest models with observational climate data and human cases to predict future levels of WNV based on future weather conditions. The Regional Model used present-day data from NY and CT, whereas the Analog Model was fit for states most closely matching the predicted future conditions in the region. Separately, we predicted changes to mosquito-based WNV risk using a trait-based thermal biology approach (Mosquito Model). The WRF model produced control simulations (present day) and pseudo-global warming simulations (future). The Regional and Analog Models predicted an overall increase in human cases of WNV under future warming. However, the Analog Model did not predict as strong of an increase in the number of human cases as the Regional Model, and predicted a decrease in cases in some counties that currently experience high numbers of WNV cases. The Mosquito Model also predicted a decrease in risk in current high-risk areas, with an overall reduction in the population-weighted relative risk (but an increase in area-weighted risk). The Mosquito Model supports the Analog Model as making more realistic predictions than the Regional Model. All three models predicted a geographic increase in WNV cases across NY and CT.