Macartney rose is a range management problem on approximately 500,000 acres of highly productive rangeland in southeast Texas. Roller chopping followed by prescribed burning is an economically effective treatment. However, there is a relatively high degree of risk associated with implementation of an effective burn. A methodology is developed to account for this uncertainty in a linear programming model designed to select optimal burning schedules. Model results indicate that consideration of uncertainty substantially affects net returns and optimal burning schedules. Low probabilities of successful burns produce law net returns and optimal burning schedules with frequent burning. Higher probabilities result in higher returns and less frequent burning.