Fluid therapy for burn shock is adjusted to establish a target level of urinary output. However, the means for adjusting infusion rate are not defined. Our objective was to compare the performance of automated computer-controlled resuscitation with manual control for burn resuscitation. Sheep with a 40% TBSA full-thickness burn, administered under halothane anesthesia, were resuscitated to restore and maintain normal sheep urinary outputs in a target range of 1 to 2 ml/kg per hour over the course of 48 hours using closed-loop resuscitation (n = 10) or manual hourly adjustment of infusion rate (n = 11). The automated closed-loop resuscitation system is based on a proportional-integral-derivative algorithm, which adjusted infusion rate based on continuous monitoring and changes in urinary output. Mean urinary outputs over the course of 48 hours were in target range and were virtually identical at 1.9 +/- 0.5 ml/kg per hour for the closed-loop group and 2.0 +/- 0.7 ml/kg per hour for the technician group. Mean infusion rates and infused volumes also were similar. The closed-loop group exhibited significantly lower hourly variation for both urinary output and infusion rate compared hourly control. Hourly targets were achieved in 41% of the measurements in technician group compared with 48% for the closed-loop group (P = .23). Hourly urinary output in the technician group was undertarget by 25% as opposed to 16% with the closed-loop group (P = .02). Automated closed-loop control of infusion rates after burn injury produced urinary outputs in target ranges with less variation and less under target values than manual hourly adjustments. Closed-loop resuscitation may provide an improvement over current resuscitation regimens.