The eddycovariance data from the heavy rainfall experiment at Feixi County in Anhui Province in 2002 is used to contrast two axisrotation method, aiming at the tilt errors of eddycovariance measurement. The results show that both triple axis rotation method and planar fit method can eliminate effectively the illusiveflux and sidestress errors caused by horizontal winds because of instrument tilt, so to improve flux precision. Both methods have its own strong points, but the planar fit method is better than triple axis rotation method relatively.