Functional analyses sometimes do not identify momentary fluctuations in the function of destructive behavior (Bowman, Fisher, Thompson, & Piazza, 1997). In such cases, individuals may mand for the reinforcer that is currently most preferred and display destructive behavior if that mand goes unreinforced. In this study, we conducted a mand analysis to test whether destructive behavior functioned as a precurrent response that increased reinforcement for the participant's mands. We then evaluated a treatment that matched this function of destructive behavior by providing differential or time‐based reinforcement of participant mands in accordance with multiple or chained schedules with reinforcement‐schedule thinning. Decreases in destructive behavior averaged 97.4% across cases. We discuss these results relative to the importance of matching treatments for destructive behavior to operant functions for both traditional and idiosyncratic functions of destructive behavior.