Abstract: The effective stress law for porous media is used extensively in multi-discipline. Numerous investigations on it have been made by previous scholars, and much knowledge about it has been obtained. However, none of universally accepted result is got from it in terms of theory all the time, for accuracy of gained effective stress coefficient is of difference with matched result in core lab. In order to get more accurate the effective stress coefficient , the maximum-likelihood function approach is applied to analysis of the experimental data of permeability from Chelmsford and Barre granite, and to comparison of them with the results obtained by conventional regression matching method and by Bernabe＇s method. The results show that maximum-likelihood function approach is feasible, with higher accuracy than that obtained by conventional regression matching method. Therefore, the effective stress coefficient determined by this approach is of higher value for reference.

Key words: permeability, stress, coefficient, maximum-likelihood function approach, pore fluid pressure