Despite the prevalence of breaks in treatments for escape-maintained problem behavior, no studies have empirically evaluated preference for different break environments (e.g., break with attention and/or tangibles). In this study, a pictorial preference assessment was conducted with 2 individuals with autism to identify preferred break environments. Assessed breaks were based on indirect assessments and direct observations. The highest- (HP) and lowest-preferred (LP) break environments and a control with no associated break were included in a reinforcer assessment using a reversal design within a concurrent-chains arrangement. Participants selected a multi-task sequence (initial link) associated with one of the break environments (terminal link). Phase A evaluated the reinforcing properties of all three breaks; the HP was removed in Phase B. Both participants allocated more responding to HP than LP, and LP than control, suggesting that breaks functioned as reinforcers. Social validity measures indicated that the results were useful for clinical programming.