In this paper, we proposed a new method for accurate optical flow estimation. The significance of this work is twofold. First in this paper we showed that the optical flow estimation could be viewed as a diffusion-reaction process (PDE-based) instead of a variational problem of minimizing some energy functions. The minimization often need the corresponding regularizer be restricted to some forms with convex and differentiable in view of convergence and the Euler-Lagrange equation. Second, we adopted coherence enhancing scheme as the diffusion part of the diffusion-reaction equation to encourage smoothing along motion boundaries and preserving them. Qualitative and quantitative results for synthetic and real-world scenes demonstrate that the new method can produce an accurate result.