To characterize the transplacental effects of cocaine on the developing brain, we have developed a mouse model of gestational cocaine exposure. Pharmacokinetic analysis revealed that cocaine and its metabolites (BE, BNE, and NC) were found in fetal brain and plasma at 30 and 120 min following SC administration to embryonic day (E) 17 pregnant Swiss Webster mice. Pregnant dams injected twice daily with cocaine HCl at 20 mg/kg SC from gestational day E8 to E17 (COC) demonstrated less food intake and lower percentage weight gain than vehicle-injected dams allowed access to food ad lib (SAL). A nutritionally paired control group of dams injected with saline vehicle and pair-fed with the COC dams (SPF) demonstrated the lowest percentage weight gain of all three groups. The surrogate fostered offspring of COC and SPF dams demonstrated persistent growth retardation [on postnatal days (P) 1, P9, and P50] and transient brain growth retardation (on P1 and P9) when compared to pups born to SAL dams. We conducted behavioral tests that allowed us to dissociate the indirect effect of cocaine-induced malnutrition from a direct effect of prenatal cocaine administration in altering postnatal behavior. Pups from all three groups were tested for first-order Pavlovian conditioning on P9 or P12, or for the ability to ignore redundant information in a blocking paradigm on P50 or P100. Unlike the SPF and SAL controls, COC mice (i.e., mice born to COC dams) were unable to acquire an aversion to an odor previously paired with shock on P9. This learning deficit was transient because on P12, COC mice trained on the same conditioning task displayed an aversion to the odor that was indistinguishable from the SPF and SAL controls. P50 and P100 COC mice (and to a lesser extent, SPF mice) demonstrated a persistent behavioral deficit in the blocking paradigm, which may reflect alterations in selective attention. We discuss how these findings in our rodent model have developmental implications for human infants exposed to cocaine in utero.