Stimulus shaping appears to be a highly successful way to teach discrimination skills. In stimulus shaping, the topographical configuration of the stimuli is gradually changed over trials so that discrimination is at first easy, and then gradually more difficult. Stimulus shaping procedures might also be effective for training visual-motor tasks. Two experiments were conducted to assess the relative effectiveness of stimulus shaping and "traditional" prompting procedures. Pegboard skills were trained in Experiment 1. In Experiment 2 a self-care skill was trained, in which children learned to hang a toothbrush or a washcloth on a specific hook. Six low-functioning retarded children were studied in each experiment, using a within-subject alternating treatments design. Each participant received concurrent training on two related tasks, using stimulus shaping for one and a standard prompting procedure for the other. Training with the stimulus shaping procedure required less training time to criterion, always resulted in fewer errors, always required fewer and less intrusive therapist's prompts, and always resulted in greater density of reinforcement. These results demonstrate the value of stimulus shaping strategies for training visual-motor skills.