A particle filtering equalization algorithm based on the optical particle sampling using the optical importance function is presented. The optical particle sampling using the optical importance function is introduced to alleviate the degeneracy phenomenon so that a re-sampling step is avoided in the method. Moreover, the proposed algorithm shares the optical importance function in J.Miguez's method, and its simplest form is derived recursively without matrix inverse. Numerical simulations show that the degeneracy phenomenon in the algorithm is alleviated greatly, and the algorithm has perfect BER performance.