A captive balloon system named InfoBalloon has been developed as a system for collecting and distributing disaster information. There is possibility of losing safety of captive balloon in the strong wind. So it is necessary to avoid operating InfoBalloon in the strong wind based on short term weather forcast and local weather forcast. This paper proposes a method of short range prediction of strong wind aloft using Kalman Filter. Then, we set an error-adjusted prediction of maximum wind speed so as to decrease operation risk. We can decide stop or resumption of InfoBalloon mooring by comparing the maximum value by the prediction with the given value of regulation. Finally, we evaluate the InfoBalloon operation rule based on the prediction method in an assumed use of captive balloon.