It is important to optimize the design of urban sewage collection pipe networks. In this study, we established a nonlinear programming mathematical model by considering the undetermined engineering system for an urban sewage collection pipe network scheme as the research subject. The model treats the minimum total annual cost of sewage collection pipe networks as the objective function, the head loss at the head and end of the branch and trunk pipes belonging to the trunk pipe section as the coupling constraint, the longitudinal slope of the pipe section, pipe economic flow rate, and optional standard pipe diameter of the division pipe section as the feasible region constraints on the decision variables, and the line scheme for the pipe networks and diameter of the division pipe section as decision variables. We developed an optimization method by combining orthogonal test-based line selection and the decomposition-dynamic programming aggregation method for the large diameter pipe system, which treats the branch and trunk pipes belonging to the trunk pipe section as the test factor, and the line scheme for the branch and trunk pipes belonging to the trunk pipe section as the test level. We considered the regional sewage collection pipe network in Taizhou city as an example to conduct a comparative analysis and verified that the proposed optimization method can solve this type of large system optimization problem rapidly and accurately, thereby providing a theoretical reference for optimizing the design of urban rainwater drainage pipe networks.