Increasing data emerging from controlled clinical trials support an analgesic activity of cannabinoids. However, the psychotropic side effects associated with tetrahydrocannabinol or synthetic derivatives essentially puts a brake on their use, possibly limiting the degree of analgesia that can be achieved as well as providing regulatory hurdles. Animal studies show that although these side effects are mediated via central cannabinoid type 1 (CB(1)) receptors, the analgesic activity in chronic pain states may be mediated via spinal CB(1) and potentially CB(2) receptors, as well as peripheral CB(1) and CB2 receptors on sensory nerves or immune cells. The design of novel compounds that either specifically target peripheral CB(1) receptors or display high selectivity for CB(2) receptors may offer avenues for harnessing the analgesic effect of CB receptor agonists while avoiding the central adverse events seen with cannabinoid structures. Clinical trials with such compounds are required to determine whether either approach can provide the level of analgesia required to fulfil the unmet medical need left by current therapies for chronic pain.