In winter, the centralized heat-supply system (CHSS) is an important and indispensable heating equipment in the northeast of China. Hence, in this article, the introduction to the CHSS is first presented, especially for the thermal dynamics of the secondary pipe network (SPN) of the CHSS, and then the identification of the thermal dynamics of the SPN under constant flow is studied. To this end, the linear parameter varying (LPV) model with autoregressive exogenous (ARX) structure is chosen as the system model and the problems of outliers and unknown system delay are both synthesized in the identification process. The physical meanings of all the variables in the LPV model are introduced in detail. By using the expectation-maximization (EM) algorithm, the equations to estimate all the parameters are derived and the desired parameter estimates can be obtained in the context. Finally, the verification tests as well as the corresponding discussions are presented.