This paper describes and validates a new experimental technique for measuring the combustion efficiency of flames in a cross flow. This technique has been developed for use in a systematic investigation of the effects of wind speed, flaring rate, and fuel composition on the performance of continuously operating flares used in the energy industry. The impact of cross flow on the overall combustion efficiency of these flames is not well understood because of the shortcomings associated with previous measurement techniques (e.g., single point aspirating probes). The proposed methodology uses a closed-loop wind tunnel to create the cross flow and to capture the products of combustion. A multiple species mass balance based on rates of accumulation of combustion products in the tunnel allows the combustion efficiency to be calculated as the fraction of carbon in the fuel being converted to CO2. Problems of leakage from the tunnel and reburning due to the recirculation of gases are accounted for in the mass balances. Two sample results of methane jet diffusion flames with combustion efficiencies of 97% and 91% are presented. These two tests are considered in detail to examine the expected statistical uncertainty and measurement sensitivities that result from this methodology. The expected uncertainty is shown to be less than 0.1% for these two cases. A sensitivity analysis shows that the most important measurements involve tracking the accumulation of CO2 in the tunnel and knowing its concentration as part of the reactant gas stream. Errors of 5% in either of these measurements produce an error in the calculated efficiency of the order of 0.4%. Hence, this methodology is shown to be an accurate and robust approach to measuring the combustion efficiency in these flows.