The high‐probability (high‐p) instructional sequence typically consists of the delivery of a series of high‐p instructions immediately followed by the delivery of a low‐probability (low‐p) instruction and is a commonly used procedure to increase compliance among children with intellectual disabilities. In the current study, we used withdrawal designs to evaluate the use of the high‐p sequence to increase compliance with a series of 2 or 3 low‐p instructions among 3 children with autism. The sequence was moderately effective in increasing compliance with all low‐p instructions for 2 participants. Compliance exhibited by the 3rd participant, for whom the high‐p sequence was ineffective, increased when differential reinforcement was delivered. We discuss the utility of the high‐p sequence to increase compliance with multiple instructions and the putative mechanisms responsible for the effects of the sequence.