We investigated the effects of a script-fading and discrimination-training procedure on mand variability in preschoolers with autism. Participants were taught to vary their vocal mands in the presence of written scripts, a green placemat, and a lag schedule of reinforcement. They were also taught to engage in repetitive mands in the presence of the same written scripts and a red placemat. When the scripts were removed, all 3 participants continued to engage in varied manding in the presence of the green placemat and lag schedule, and they continued to engage in repetitive manding in the presence of the red placemat. When the lag schedule was also removed, 2 of the 3 participants continued to engage in varied responding in the presence of the green placemat and repetitive responding in the presence of the red placemat. Finally, all 3 participants demonstrated generalization and maintenance of mand variability during snack sessions with their peers.