Growth asynchrony in trees increases uncertainty in modeling forest productivity and ecological services. Despite recognition of growth variability among trees, the process of asynchronous growth and its ecological implications are poorly understood. We used tree-ring data obtained from increment core samples in 1046 juniper trees at 32 sites on the Tibetan Plateau and 538 pine trees at 20 sites in eastern China to explore patterns of growth asynchrony in history. The results showed that, although ring-width series are correlated among trees, there exists widespread growth asynchrony among trees at inter-annual scale. After highly synchronous growth events, the number of trees keeping continuous growth synchrony declined progressively and, notably, in a relatively uniform exponential function (y = b ^x-1), with parameter b ranging from 0.57 to 0.83 for juniper forests on Tibetan Plateau and ranging from 0.56 to 0.72 for pine forests in eastern China. Forest recovery and resilience to climatic extremes were higher in years of low growth synchrony than years of high growth synchrony. We argue that de-synchronization in tree growth spreads the capacity of trees to respond to disturbances and stresses at the individual level, consequently enhances forest resilience to risks of climate uncertainty. This resilience-acquisition strategy is an inherent and an adaptive property of forest tree growth and application of the strategy in forest modelling will help assessment of forest health risks under future climate change.