Cocaine use disorder is a serious public health issue for which no approved pharmacotherapies exist. The development of a pharmacotherapy for cocaine use disorder is a priority for the National Institute on Drug Abuse. Amphetamine maintenance has been shown to be effective to reduce cocaine use in double-blind placebo controlled clinical trials, but has not been approved due to concerns over safety and abuse liability. Development of new pharmacotherapies is facilitated by preclinical testing for effectiveness and identification of new targets for medication development. The first part of this dissertation develops a novel non-human primate cocaine self-administration choice procedure that is modeled after a human laboratory cocaine self-administration choice procedure to improve translational research and facilitate medication development. The second part of this dissertation is devoted to examining the mechanisms of amphetamine maintenance-induced decreases in cocaine use. In the novel non-human primate choice procedure, monkeys chose between injections of cocaine or food pellets (0, 1, 3 or 10) in a 9-choice discrete trials procedure. The reinforcers were available on concurrent independent progressive-ratio schedules. Monkeys chose between cocaine and food in a dose- and magnitude-dependent manner. Maintenance on 7 days of lisdexamfetamine and amphetamine decreased cocaine choices without decreasing food responding, providing evidence that this model may be able to predict drugs that will have clinical efficacy to decrease cocaine use. The next set of experiments examined the effects of amphetamine maintenance on the abuse-related behavioral (intracranial self-stimulation, ICSS) and neurochemical [nucleus accumbens dopamine (DA) and serotonin (5-HT)] effects of cocaine, methylenedioxypyrovalerone, and methamphetamine in rats. Amphetamine maintenance produced sustained increases in ICSS baseline responding and nucleus accumbens DA levels without affecting 5-HT levels. Amphetamine maintenance also attenuated the behavioral and neurochemical abuse-related effects of cocaine but not those of methamphetamine, and with MDPV, amphetamine maintenance decreased the abuse-related neurochemical effect of MDPV, but not the abuse-related behavioral effect. This suggests that amphetamine would likely be most effective against cocaine, least effective against methamphetamine and between the two for MDPV. These data suggest targets that selectively release DA will be the most effective against cocaine use disorder.