This study aimed to examine the effectiveness of a prompting procedure (using a textual script without fading) in increasing novel behavior during conversations by children with autism. Four children with autism participated in the study. Dependent variables were scripted behavior, unscripted behavior, novel behavior, and inappropriate statements in a conversation between children with autism. This study used an ABCA design, the baseline, a script procedure, a script removal, and a return to baseline phase. In the baseline, participants infrequently emitted novel behavior and emitted inappropriate statements. When the script procedure was introduced, all participants emitted novel behavior and appropriate behavior more often than during baseline with fewer inappropriate statements. These gains were maintained in the script removal and return to baseline conditions. Three out of four participants continued to emit novel behavior during follow up probes. These results indicate that the script procedure without fading was useful for increasing novel behavior and appropriate statements.