The purpose of this work is to find the probabilistic-time characteristics of network service during the transmission of multimodal information. For this purpose, a mathematical model of a heterogeneous multimodal data transmission system has been proposed, the peculiarity of which is to consider Markov modulated Poisson flows as input modality flows, each of which is controlled by one Markov chain with a finite number of states, specified by a matrix of infinitesimal characteristics. Each input modality of a certain type is served for a random time according to the exponential probability distribution law. The constructed model was studied numerically and conclusions were drawn about promising directions for further research.