The resistance of forests to extreme climatic events such as drought shapes their sensitivity to future extreme events in space and time. To a large extent, the ability of trees to learn from prior droughts explains how trees adjust their sensitivity to water deficit. We use tree-ring width data collected from 1565 juniper trees (Juniperus prezwalskii and Juniperus tibetica) across 57 sites on the Tibetan Plateau to model tree resistance to water deficit and to map drought sensitivity across the species’ distribution. We test the effect of both the frequency and severity of drought on the drought memory of trees. We find that trees at mid-latitudes and in the northwestern part of the juniper distribution range exhibit higher drought sensitivity. Water deficit is the main factor controlling tree resistance and thus affects spatial sensitivity to drought. At wetter sites, higher drought frequency enhances tree adaptability through ecological memory, thereby promoting tree resistance and decreasing tree sensitivity to extreme drought events. At drier sites, higher drought frequency causes junipers’ growth decline that is not beneficial for tolerance to extreme drought events. Regional drought conditions and the frequency of pre-droughts affect tree resistance and sensitivity to extreme drought events. This explains the spatial pattern of drought risk for juniper forests on the Tibetan Plateau, and helps us to better understand the vulnerability of this high-elevation forest ecosystem. Such information is important for maintaining forest health and informing the sustainable development of the Tibetan Plateau under a changing climate.