Interrupted time-series (ITS) is a quasi-experimental design which evaluates the effectiveness of an intervention based on time-series outcome variables. Compared with the single group of ITS, the two groups of ITS can better control the influence of pre-interventional confounding factors and evaluate the effectiveness of the intervention. This paper summarizes the principles and statistical methods of two groups of ITS by an example of evaluating vaccine effect on the incidence of a disease in two cities. The regression model is fitted by Prais-Winsten method and Newey-West method and the results are explained and compared in detail. When the intervention is performed with other confounding interventions at the same time, the two groups of ITS can be more effective to balance the existing trends before the intervention, and evaluate the effectiveness of intervention. The method of two groups of ITS has important practical significance, providing new insights in program evaluation.