The trial included 24 children (aged 2–7 yr) referred for dental treatment under general anesthesia, since conventional behavioral management methods had failed to achieve treatment acceptance. As an alternative, they received, on two separate occasions with “identical” dental treatment, conscious sedation by rectal administration of either midazolam (0.3 mg/kg body weight (bwt)) or midazolam (0.3 mg/kg bwt) plus ketamine (1.0 mg/kg bwt). This allowed a double‐blind, crossover design. The aims were to assess conscious sedation, combined with local anesthesia, as an alternative to general anesthesia, and further to evaluate the effects obtained by addition of a low dose of ketamine to rectally administered midazolam. The feasibility of dental treatment was rated as excellent or good for 16 of the 24 children when premedicated with midazolam, and for 18 of the 24 children when ketamine was added to midazolam. At least some treatment could be given to all children. Verbal contact was maintained with all children throughout both treatment sessions. The children were significantly less anxious when they arrived for the second session. Amnesia and drowsiness were significantly increased when ketamine was added to midazolam. The combination also tended to be more efficient in relief of anxiety and prevention of pain, but there were large variations in the children's responses to the drugs. Midazolam significantly reduced the blood oxygen level, but not with ketamine added. For most children, both regimens proved to be appropriate as alternatives to general anesthesia. From a pharmacologic point of view, the combination of midazolam and ketamine appears to be reasonable because 1) both drugs have sedative and amnestic properties, 2) ketamine adds an analgesic component, 3) midazolam counteracts the psychic side‐effects of ketamine, and 4) ketamine counteracts the depressive effects of midazolam on vital body functions (respiration and circulation).