The knowledge of the strain/stress field evolution in time is fundamental to the understanding of the dynamic system. Based on the principle that past tectonic stress should have left traces in the rocks, geologists have been trying to determine the paleostress history from evidence found in rocks for decades. Recent development of techniques for automatic extraction of fracture surfaces from digital outcrop models and estimation of historical shear deformation on rock fractures provide an efficient way of quantitatively acquiring large amounts of high‐quality fracture/fault slip data from outcrops. Unlike traditional paleostress inversion methods whose data are manually collected in the field, the new techniques provide much more detailed information about the strain of the outcrop and a good opportunity to develop quantitative methods for deciphering more realistic paleostrains. In this study, instead of fitting the slip data from several fractures to calculate the overall strain tensor, the local strain tensor is calculated for slip on each fracture from the outcrop. Then, the local strain tensors are grouped into populations corresponding to different strain events using a clustering analysis technique. Theoretical advantages of this new method over the traditional ones are discussed. The applications on outcrops in the eastern Tian Shan area give a clear picture of the late Cenozoic paleostrain variation over space and time and also throw light on the cause for the change in the strain regime in time, the fracture development patterns, and the distribution of shear displacements in fracture networks.