Abstract: The paper constructed one of transition probability matrix based Markov chain theory and random theory in discrete time,and we divide the state of traffic parameters of communication into N levels according to characteristic of communication traffic parameters that will be forecasted. Therefore, we get a finite state space S=1,2,…, N. As a result,a forecasting model about traffic parameters of communication was established with transition probablilty matrix and random theory,and the reliablilty of the model was verified through two real examples on communication networks.