Background and Purpose: Maintenance treatment with opioid agonists (buprenorphine, methadone) decreases opioid use and relapse. We recently modelled maintenance treatment in rats and found that chronic delivery of buprenorphine or the μ opioid receptor partial agonist TRV130 decreased relapse to oxycodone seeking and taking. Here, we tested the buprenorphine analogue BU08028 on different heroin relapse-related measures and heroin versus food choice. Experimental Approach: For relapse assessment, we trained male and female rats to self-administer heroin (6 h·day−1, 14 days) in Context A and then implanted osmotic minipumps containing BU08028 (0, 0.03 or 0.1 mg·kg−1·d−1). Effects of chronic BU08028 delivery were tested on (1) incubation of heroin-seeking in a non-drug Context B, (2) extinction responding reinforced by heroin-associated discrete cues in Context B, (3) reinstatement of heroin-seeking induced by re-exposure to Context A and (4) re-acquisition of heroin self-administration in Context A. For choice assessment, we tested the effect of chronic BU08028 delivery on heroin versus food choice. Key Results: Chronic BU08028 delivery decreased incubation of heroin seeking. Unexpectedly, BU08028 increased re-acquisition of heroin self-administration selectively in females. Chronic BU08028 had minimal effects on context-induced reinstatement and heroin versus food choice in both sexes. Finally, exploratory post hoc analyses suggest that BU08028 decreased extinction responding selectively in males. Conclusions and Implications: Chronic BU08028 delivery had both beneficial and detrimental, sex-dependent, effects on different triggers of heroin relapse and minimal effects on heroin choice in both sexes. Results suggest that BU08028 would not be an effective opioid maintenance treatment in humans.