Mechanical design activities can be categorized as two classes: creating new designs for new problems and modifying old designs to fit new problems. A vast majority of mechanical design activities can be associated with the latter class. In most cases, it is more effective to modify the design process that created the mechanical artifact than it is to modify the mechanical artifact itself. A scheme for design automation that uses Analogical Problem Solving (APS) as its intelligent agent offers an effective method to solve both classes of mechanical design. It is particularly suited to the latter class of design activities. It would rely heavily on a design knowledge base for storing design cases generated while solving design problems, and retrieving design cases that are applicable in a new design problem context. In this paper, a cognitive model of memory for storing design plans is presented. The memory model is four layers deep: product design plans, assembly design plans, mechanical component design plans, and recurring engineering problems. The storage and retrieval mechanism is based on some of the more popular work found in Case Based Reasoning. To alleviate the user(s) from operating in a restricted vocabulary set, a semantic network is integrated into the memory model for use as an elaboration and cross-reference mechanism. Mechanical design plans for products, assemblies, mechanical components and recurring engineering problems can be stored and retrieved from the memory model using information found in the description of the design problem. Sample examples are presented to demonstrate the potential of the memory model.