In the process of product development, there is a complex information-dependent relation between coupling design tasks with single input and multiple output. Because of the large amount of executed tasks and transmitted information, the complicated relationship between these tasks is hard to be sorted out by using the single-stage task execution method, which may lead to problems such as the unreasonable allocation of tasks in the process of task execution, excessive iteration rework times between tasks and too long product development cycle. To solve these problems, an information processing strategy for the task iteration was put forward, and a two-stage iteration model and the mathematical model of calculating task execution time were built based on the deep analysis of a single-stage iteration model for the coupling design task with single input and multiple output. The best two-stage task allocation scheme was obtained by working out this mathematical model. The effectiveness of this method was proved through the example of an engine development project. The results showed that the iterative execution time of tasks was shorter than single-stage by using the two-stage iteration model and the shortest iterative execution time of the two-stage iteration model could be obtained by reasonably assigning tasks into two phases. The method selects the appropriate task iterative model according to the different information relations between tasks, which can effectively shorten the product development cycle and provide some theoretical reference for product designers to choose reasonable task execution way in the process of product development.