Summary: In this paper, a multigroup SEIR epidemic model with vaccination is firstly introduced and its basic reproduction number is derived for epidemiological prediction. An algorithm to generate complex network representing heterogeneous structure of individuals is heuristically proposed. The network degree observed from the presented algorithm is normally distributed at significance level 0.05. Community detection method based on genetic algorithm for optimizing network modularity is used to partition individuals into appropriate communities. In numerical simulation on complex network, the disease transmissions within community and across communities are considered. Tukey’s HSD test indicates that different vaccination levels have significant effect on number of infected cases at significance level 0.05.