The soaring demand, inevitable changes, and substantial change costs associated with complex mechanical products (CMPs) have accelerated the need to reasonably and accurately respond to changes in customer requirements during the product design process. However, current related studies cannot provide a simple and intuitive decision reference for decision-makers (DMs) to respond to these changes. In this work, a complex network theory-based methodology is proposed. First, a complex network model of CMPs is constructed; this model is processed unidirectionally through analysis of the constraint relation and affiliation among parts. Second, all nodes in this network are divided into levels, and all feasible change propagation paths are selected by breadth-first search. Furthermore, to quantify the change losses of paths, a novel “change workload” is proposed, which is a comprehensive indicator, and a distinct decision reference. The “change workload” is composed of “network change rate,” “change magnification node rate,” and “change magnification rate,” whose weights are evaluated by The Entropy Method and Technique for Order Preference by Similarity to an Ideal Solution. Due to the independence of the “change workload” from expert experience, the proposed methodology is reasonable and capable of outputting a list of affected parts and a preferred ordering of propagation paths, which could provide clearer and more direct guidance for DMs. This presented method is fully proven through a real-world case study of a wind turbine.