In this paper we propose a hierarchical deformation model to describe the deformation of cursive Chinese characters for on-line character recognition. The new approach consists of two levels of match processes. First, the attributed string editing algorithm determines the stroke correspondence between the input and the reference patters. Next, the constrained parabola transformation is used to reduce the difference between the matched strokes appropriately. Experimental results show that our hierarchical deformation model provides a robust distance measure between patterns, and obtains a quite accurate approximation to the deformation of cursive Chinese characters with much lower computational cost.