This study aimed to teach individuals with autism spectrum disorder (ASD) and limited vocal speech to emit target vocalizations while using a speech-generating device (SGD). Of the 4 participants, 3 began emitting vocal word approximations with SGD responses after vocal instructional methods (delays, differential reinforcement, prompting) were introduced. Two participants met mastery criterion with a reinforcer delay and differential reinforcement, and 1 met criterion after fading an echoic model and prompt delay. For these participants, vocalizations initiated before speech outputs were shown to increase, and vocalizations generalized to a context in which the SGD was absent. The 4th participant showed high vocalization rates only when prompted. The results suggest that adding vocal instruction to an SGD-based intervention can increase vocalizations emitted along with SGD responses for some individuals with ASD.