OBJECTIVE Methods to promote knee pain analgesia without impairing motor function have been widely researched. This study aimed at comparing intra-articular morphine, bupivacaine, and fentanyl analgesic effects (as compared to saline solution), after knee video-arthroscopy. METHODS Participated in this study 60 patients who were randomly distributed in four groups: GI (n = 15) 10 ml saline solution; GII (n = 15) 2 mg morphine diluted in 10 ml saline solution; GIII (n = 15) 10 ml of 0.25% bupivacaine; GIV (n = 15) 100 microg fentanyl diluted in 10 ml saline solution, injected at surgery completion. All patients received spinal anesthesia with 15 mg hyperbaric bupivacaine. Pain intensity was evaluated by a visual analog scale (VAS) (at surgery completion, and 6, 12, 18 and 24 hours later). The need for analgesic complementation (1 g intravenous dipirone) was also evaluated. Side effects were recorded. RESULTS There were no statistical differences in pain intensity among groups in almost all moments studied. There was a statistic difference up to 6 hours, when the fentanyl group had significantly lower pain. The morphine group needed more dipirone complementation. Side effects were minor, without statistical significance. CONCLUSIONS There were no statistical differences among solutions analgesic effects in almost all moments studied.