Neuropeptide S (NPS) regulates various biological functions by selectively activating the NPS receptor (NPSR). Recently, epidemiological studies revealed an association between NPSR single nucleotide polymorphisms and susceptibility to panic disorders. Here we investigated the effects of NPS in mice subjected to the elevated T maze (ETM), an assay which has been proposed to model anxiety and panic. Diazepam [1 mg/kg, intraperitoneally (i.p.)] elicited clear anxiolytic effects reducing the latency to emerge from the closed to the open (CO) arm without modifying the latencies from the open to the closed (OC) arm. By contrast, chronic fluoxetine (10 mg/kg i.p., once a day for 21 days) selectively increased OC latency, suggesting a panicolytic-like effect. NPS given intracerebroventricularly at 0.001–1 nmol elicited both anxiolytic- and panicolytic-like effects. However, although the NPS anxiolytic dose–response curve displayed the classical sigmoidal shape, the dose–response curve of the putative panicolytic-like effect was bell shaped with peak effect at 0.01 nmol. The behaviour of wild-type [NPSR(+/+)] and receptor knock out [NPSR(−/−)] mice in the ETM task was superimposable. NPS at 0.01 nmol elicited anxiolytic- and panicolytic-like effects in NPSR(+/+) but not in NPSR(−/−) mice. In conclusion, this study demonstrated that NPS, via selective activation of the NPSR, promotes both anxiolytic- and panicolytic-like actions in the mouse ETM.