Psychedelics comprise drugs come from various pharmacological classes including 5-HT 2A agonists, indirect 5-HT agonists, e.g., MDMA, NMDA antagonists and κ-opioid receptor agonists. There is resurgence in developing psychedelics to treat psychiatric disorders with high unmet clinical need. Many, but not all, psychedelics are schedule 1 controlled drugs (CDs), i.e., no approved medical use. For existing psychedelics in development, regulatory approval will require a move from schedule 1 to a CD schedule for drugs with medical use, i.e., schedules 2–5. Although abuse of the psychedelics is well documented, a systematic preclinical and clinical evaluation of the risks they pose in a medical-use setting does not exist. We describe the non-clinical tests required for a regulatory evaluation of abuse/dependence risks, i.e., drug-discrimination, intravenous self-administration and physical dependence liability. A synopsis of the existing data for the various types of psychedelics is provided and we describe our findings with psychedelic drugs in these models. FDA recently issued its guidance on abuse/dependence evaluation of drug-candidates (CDER/FDA, 2017). We critically review the guidance, discuss the impact this document will have on non-clinical abuse/dependence testing, and offer advice on how non-clinical abuse/dependence experiments can be designed to meet not only the expectations of FDA, but also other regulatory agencies. Finally, we offer views on how these non-clinical tests can be refined to provide more meaningful information to aid the assessment of the risks posed by CNS drug-candidates for abuse and physical dependence. This article is part of the Special Issue entitled ‘Psychedelics: New Doors, Altered Perceptions’.