We analyze climate simulations conducted with the RegCM3 regional climate model on 50‐ and 15‐km model grids to diagnose the dependence of wildfire incidence and area burned variations on monthly climate long‐term means and anomalies over North America for the period 1986‐2013. We created a new wildfire data base by merging the Fire Program Analysis Fire‐Occurrence Database, the National Interagency Fire Center Fire History Data (NIFC), and the Canadian National Fire Database. The data base includes 2,083,865 daily fire starts that burned a total of 1.25×108 ha in North America. We derive long‐term climatologies, standardized gamma indices, and composite climate anomalies of atmospheric circulation (500 hPa height and wind) and various surface fields (e.g., solar radiation, soil moisture, vapor pressure deficit, latent and sensible heat fluxes) to illustrate the climatology of burned area. The immediate and lagged monthly atmospheric circulation and surface climate anomalies differentiate high and low fire years, and the role of ENSO in wildfire occurrence. Our approach demonstrates the association of the seasonal cycles of wildfire and climate and the strong role of climatic variability in modulating the seasonal cycle as a control of wildfire on monthly time scales.