The assessment of drug effects on attention is important in non-clinical pharmacology, for both evaluation of safety and therapeutic efficacy of medicinal products. In the present study, we have developed a two-lever choice behavioral test to assess drug effects on attentional performance in monkeys. In each trial of this experiment, one of two lamps in front of a monkey was randomly illuminated for a brief period of time and the monkey was required to press a lever beneath the lamp 30 times to obtain a food reward. The percentage of correct responses, response latency of correct choice responses and response speed were measured. Using this test, we examined the effects of three sedative drugs, diazepam (0.25, 1 and 4 mg/kg, i.g.), ethanol (0.5, 1 and 2 g/kg, i.g.), and pentobarbital (0.25, 1 and 4 mg/kg, i.v.). Diazepam and pentobarbital lengthened response latency without significantly affecting the percentage of correct responses, response and response speed, suggesting selective disruptive effects on attentional performance. In contrast, ethanol at the high dose tested caused deterioration in all three measurements, which is thought to reflect a general sedative effect including motor impairment as reflected by lengthening response speed. It is suggested that the present behavioral test method could detect drug effects on attentional performance in monkeys and could be a useful tool for safety assessment in drug development.