Selective cyclooxygenase-2 non-steroidal anti-inflammatory drugs are known to inhibit bone repair, especially when long-term administration is required due to chronicle inflammatory diseases. In order to evaluate the action of this drug in bone repair during short-term administration, 48 rats underwent surgical bone defects in their tibias, being randomly distributed into three groups: (Group 1) negative control; (Group 2) animals treated with celecoxib, and (Group 3) animals treated with ketoprofen, both experimental groups at 1 mg/kg dose, beginning 1 h before the surgical procedure and after every 12 h for the following 3 days, or until the day of sacrifice. The animals were killed after 48 h, 7, 14, and 21 days. The tibias were removed for morphological, morphometric, and immunohistochemistry analysis for COX-2. No statistical significant differences were observed in the quality of bone repair and quantity of formed bone among the groups. COX-2 immunoreactivity of the celecoxib treated specimens was more intense in the first analyzed period, and no longer observed in the periods of 14 and 21 days. Such results suggest that the administration of the analyzed drugs in short periods does not interfere with the process of bone repair in the tibia of rats.