Delay discounting (DD) refers to how rapidly an individual devalues goods based on delays to receipt. DD usually is considered a trait variable but can be state-dependent; yet few studies have assessed commodity valuation at short, naturalistically relevant time intervals that might enable state-dependent analysis. This study aimed to determine whether drug-use impulsivity and intelligence influence heroin DD at short (ecologically relevant) delays during two pharmacological states (heroin satiation and withdrawal). Out-of-treatment, intensive heroin users (n=170; 53.5% African-American; 66.7% male) provided complete DD data during imagined heroin satiation and withdrawal. Delays were 3, 6, 12, 24, 48, 72, and 96 hours; maximum delayed heroin amount was thirty $10 bags. Indifference points were used to calculate area under the curve (AUC). We also assessed drug-use impulsivity (subscales from the Impulsive Relapse Questionnaire, IRQ) and estimated intelligence (Shipley IQ) as predictors of DD. Heroin discounting was greater (smaller AUC) during withdrawal than satiation. In regression analyses, lower intelligence and IRQ Capacity for Delay as well as higher IRQ Speed (to return to drug use) predicted greater heroin discounting in the satiation condition. Lower intelligence and higher IRQ Speed predicted greater discounting in the withdrawal condition. Sex, race, substance use variables, and other IRQ subscales were not significantly related to the withdrawal or satiation DD behavior. In summary, heroin discounting was temporally rapid, pharmacological state-dependent, and predicted by drug-use impulsivity and estimated intelligence. These findings highlight a novel and sensitive measure of acute DD that is easy to administer.