Traditional design methods based on mathematical formulations often fail to make full use of knowledge specific to the design of noncontinuous plants. In this paper we present an evolutionary design method which solves complex design problems by systematic and comprehensive exploitation of such knowledge. Through a graphic user interface, the approach allows the design in progress to be interrupted to incorporate human judgment.The evolutionary design approach begins with an initial design containing a minimum number of equipment units, and modifies the design to make it feasible with the addition of as few units as possible. A sizing procedure is then used to minimize the total equipment cost. Both synthesis and sizing use prediction models and decision models which we have developed specifically for noncontinuous plants. With the proposed method, we are able to solve design problems that are too complicated for other design methods. Our method is illustrated by two complete examples.