Aiming at Hidden Markov model’s disadvantages in image denoising，parameter solving method is used. Firstly for the parametric model of ternary group its restrictions are determined， state probability is computed with recursive function，expectation maximization is made by using maximum likelihood estimation. The process includes expectation and maximization. In image denoising，its threshold is set through Kullback-Leibler distance. Finally the parameter solution is given. Simulation results show that the proposed algorithm can keep the useful information of the image，and computation speed is fast.