In this article, three corrections are applied to improve the simultaneous optimization synthesis approach (SOSA) of heat exchanger networks (HENs). First, the temperature-dependent characteristics of process streams are considered in calculating their heat capacities, which are regarded as constant within a defined temperature interval; second, splitters and mixers of streams located in the network are listed as investment items, resulting in more detailed HEN cost calculations; third, a group of heuristic rules is suggested to cancel superstructure matches that are impossible in terms of engineering or thermodynamics before the optimization calculation, making the simultaneous optimizations on both investment cost and utility consumption charge available while also saving significant computation time. Based on these corrections, an improved SOSA is developed in which an MINLP model is solved by the use of genetic algorithms to avoid being trapped in a local optimum. A case study shows that this approach realizes significant progress in both precision and computation-time savings.