A class of iterative aggregation/disaggregation methods for computing some important characteristic of Markov chain such as stationary probability vectors and mean first passage times matrices is presented.