Chronic drug abuse is a complex behavioral and social phenomenon, that stems from a diverse set of underlying neural mechanisms. However, two defining features of drug addiction make it especially difficult to treat. First, addiction is compulsive —individuals often continue or resume drug use despite a conscious, stated wish to quit. Second, it is persistent —relapse to active drug use can occur despite years of abstinence. This chapter discusses evidence that the inappropriate engagement of neural mechanisms involved in normal associative learning is responsible for these key behavioral aspects of drug addiction. Central to understanding these mechanisms are the distinctions between reward and reinforcement, between goal-directed and automatized behavior, and between information-dense patterns of synaptic plasticity and information-poor neuronal adaptations. This chapter is intended to complement a previous review (1 ) by elaborating on some particular forms of associative learning that may be central to drug addiction, although limitations of space prevent a full consideration of the many roles that learning has been suggested to play in drug addiction (2 –5 ).