Regulating methanol feed concentration in direct methanol fuel cells (DMFCs) is important for improving electrical performance and fuel utilization. Low methanol concentration reduces the reaction rate at the anode due to Nernstian effects resulting in a lower operating voltage. However, simply increasing the methanol concentration does not always lead to improved performance due to increased methanol crossover from the anode to the cathode resulting in mixed-potential losses and the associated fuel loss. Hence, there exists an optimal intermediate value of methanol concentration for each current density that will yield the highest electrical performance (V). In this paper, we describe the development of an in situ methodology which uses the measured cell voltage as the feedback to regulate the methanol feed concentration for maximum power density. This methodology is demonstrated at the current densities of 50, 100, and 250 mA cm(-2) and the results for optimal concentration are presented. Fuel loss as a function of methanol concentration is evaluated by oxidizing the crossover methanol at the cathode exhaust and measuring the CO2 mass flux. (C) 2008 Elsevier B.V. All rights reserved.