•  While cholinergic interneurones are few in number, their axons densely innervate and provide abundant acetylcholine‐containing terminals in the nucleus accumbens (NAc). •  Although cholinergic effects on inhibitory postsynaptic currents (IPSCs) via nicotinic and muscarinic receptors have been reported, the relationship between cholinergic modulation of IPSCs and presynaptic cell subtypes in the NAc has remained elusive. •  Here we show that muscarinic receptor activation suppresses IPSCs in medium spiny neurone (MSN)→MSN connections, whereas nicotinic receptor activation enhances IPSCs in fast‐spiking neurone→MSN connections. •  These reciprocal regulatory mechanisms for IPSCs help us to understand the role of cholinergic processes in physiological and pathophysiological functions of the NAc. Abstract  Medium spiny neurones (MSNs) in the nucleus accumbens (NAc) are the principal neurones whose activities are regulated by GABAergic inputs from MSNs and fast‐spiking interneurones (FSNs). Cholinergic interneurones play important roles in the regulation of activity in MSNs; however, how acetylcholine modulates inhibitory synaptic transmission from MSNs/FSNs to MSNs remains unknown. We performed paired whole‐cell patch‐clamp recordings from MSNs and FSNs in rat NAc shell slice preparations and examined cholinergic effects on unitary inhibitory postsynaptic currents (uIPSCs). Carbachol (1 μm) suppressed uIPSC amplitude by 58.3 ± 8.0% in MSN→MSN connections, accompanied by increases in paired‐pulse ratio and failure rate, suggesting that acetylcholine reduces the probability of GABA release from the synaptic terminals of MSNs. Carbachol‐induced uIPSC suppression was antagonised by 100 μm atropine, and was mimicked by pilocarpine (1 μm) and acetylcholine (1 μm) but not nicotine (1 μm). Application of AM251 slightly reduced carbachol‐induced uIPSC suppression (30.8 ± 8.9%), suggesting an involvement of endocannabinoid signalling in muscarinic suppression of uIPSCs. In contrast, FSN→MSN connections showed that pilocarpine had little effect on the uIPSC amplitude, whereas both nicotine and acetylcholine facilitated uIPSC amplitude, with decreases in failure rate and paired‐pulse ratio, suggesting that nicotine‐induced uIPSC facilitation is mediated by presynaptic mechanisms. Miniature IPSC recordings support these hypotheses of presynaptic cholinergic mechanisms. These results suggest a differential role for muscarinic and nicotinic receptors in GABA release, which depends on presynaptic neuronal subtypes in the NAc shell.