Repetitive behavior and delays in communication are core deficits of autism spectrum disorder. As a result, individuals with autism often engage in repetitive verbal behavior, and they may not vary their verbal behavior, even when the situation demands it. The purpose of this study was to investigate the effects of a script training 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 Lag schedule of reinforcement. They were also taught to not vary their vocal mands in the presence of the same written scripts and a red placemat. When the scripts were removed, all three participants continued to engage in varied manding in the presence of the Lag schedule of reinforcement and the green placemat. All three participants also did not vary their mands in the presence of the red placemat. When the Lag schedule of reinforcement was removed, two participants continued to engage in varied responding in the presence of the green placemat and unvaried responding in the iv presence of the red placemat. One participant did not engage in varied responding when the Lag schedule of reinforcement was removed. However, when the Lag schedule of reinforcement was re-introduced, varied responding re-emerged. Finally, all three participants demonstrated mand variability during snack sessions when their peers were present, and they maintained their varied manding after a 2-week follow-up. Therefore, the purpose of this study was to determine whether or not three preschoolers with autism would engage in varied or repetitive requests, depending on the situation.Participants were taught to respond in varied or repetitive requests in the presence of green and red placemats, respectively. Written sentences were also available on each placemat to help participants engage in independent requests. When the written sentences were removed, all three participants continued to engage in varied requests in the presence of the green placemat and when reinforcement was provided for response variability; repetitive requests continued in the presence of the red placemat. Two out of three participants continued to engage in varied requests in the presence of the green placemat when reinforcement was provided for all responses, regardless of whether or not they were varied. Finally, all three participants engaged in varied requests during their typical preschool snack session, and they also continue to vary their requests after a 2-week follow-up. Our results indicate that preschoolers with autism can vary the way they make requests, depending on the context.