This paper presents a heuristic procedure for designing paced production lines with stochastic task times. The procedure adopts explicitly the objective of minimizing the sum of incompletion cost and labour cost. Basically, the first stage of the procedure uses a probabilistic procedure to generate many promising line designs: this probabilistic procedure is conceptually related to the procedures of Arues (1960) and Tongc (1965). The procedures second stage uses the Kottas-Lau (1976) cost model to evaluate the operating costs of the generated line designs and to identify the economical ones. Extensive tests with the computerized procedure using randomly generated-test problems have indicated that the procedure can produce very good line designs with modest computer time