Classical conditioning is widely used to study motivational properties of addictive drugs in animals, but has rarely been used in humans. We established a procedure suitable for studying the neurobiology and individual determinants of classical conditioning in humans. Healthy volunteers were randomly assigned to four groups that received methamphetamine or placebo in the presence of distinctive environmental cues under paired or unpaired conditions. During each session, subjects performed tasks known to activate the ventral striatum. Tasks were performed in the presence of a distinctive context, consisting of a screen background image of a beach or mountains, accompanied by corresponding sounds. Separate groups of subjects carried out the tasks under high ($35-50) or low ($5-20) reward conditions. Within each of the two reward conditions, one group (paired) received methamphetamine (20 mg, oral) or placebo consistently associated with one of the contexts, while the other (unpaired) received drug or placebo unrelated to context. A fifth group (paired) performed the tasks with contextual cues but in the absence of monetary incentives. Before and after conditioning, participants carried out a series of forced choice tasks for the contextual cues, and change of preference over time was analyzed. All paired groups showed a significant increase in preference for the drug-associated context, with a linear trend for increase across the levels of reward. Preference was unrelated to subjective drug effects, and did not change in the unpaired group. These data support the translational utility of our conditioning procedure for studies of reward mechanisms in humans.