Sutong Changjiang Highway Bridge is the longest cables stayed bridge in the world at present. The correct design wind velocity is highly important to design, construction and operation of the bridge. In order to obtain the reasonable design wind velocity of the bridge, four wind observations are established both on the bank and on the bridge reach of the Changjiang River and both in Nantong and Changshu meteorological observation stations. Further more, a wind observation tower of 80 m in height is also set up on the bank at the same place to measure gradient wind, that is the wind on 10 m, 30 m, 50 m, 70 m, 80 m in height respectively. Wind simultaneous observations in four stations above mentioned and on wind observation tower are then in process from 1st March 2000 to 28th February 2003 to collect the basic data which are necessary for calculating the design wind velocity of the bridge. On the basis of contrasting the corresponding wind data from the four wind observing stations, long series history wind data from nearby meteorological observation station are extended into the bridge reach on the river. By using the gradient data, the variation law of wind velocity with height is studied. As a result, the fitting method of frequency distribution of extreme value is used here to calculate and estimate the reasonable basic designing and engineering wind velocity, which are requisite to the bridge construction. Finally, five conclusions are found as follows.(1) Because of the influences caused by the different geographic environments, the wind velocity of the bridge reach is obviously stronger than that of the nearby meteorological stations, especially the velocities of easterly and northwesterly wind. Generally speaking, the wind velocity of the bridge reach on Changjiang River is about twice as strong as that of the nearby meteorological stations on continent.(2) The wind velocity of the bridge reach on the river surface is also evidently stronger than that on the bank of the river, so that in practical application the wind velocity on the bank can not be substituted for that on the river surface.(3) There is a fine linear correlation relationship between the velocities of the bridge reach and of the nearby meteorological observation stations. When calculating the velocities of the bridge reach, results obtained by using the data of two nearby stations, i.e., Nantong and Changshu observatories, are better than those gained by using the data of one nearby station only. After that, according to insurance probability the confidence interval of design wind velocity is determined.(4) Variation of the wind velocities of the bridge reach with height is in keeping with the exponential law. Here the exponent value is equal to 0.118 which is less than recommended by standards.(5) The wind velocity samples of equal and more than 6 m/s are chosen to compute the basic design wind velocity of Sutong Changjiang Highway Bridge. Though the result of 38.9 m/s is slightly stronger than that simply calculated in the light of standards, the bridge can resist the wind disaster which has return period of 100 years.Therefore, the reliable scientific basis of resistance against wind disaster is provided to making designs for Sutong Changjiang Highway Bridge by the research results of this paper. What is more, the new method of calculating the wind velocity on the river surface developed in the paper may be popularized to other water surface.