This paper presents a practical design for hybrid adaptive control which does not need any numerical integration and is not destabilized by computational delays for parameter estimation. The controller consists of two simultaneously executing parts, that is, a control part and an estimation part. The estimation part searches the maximum estimation error over a successive time interval repeatedly. Based on this maximum error, the estimators for the unknown parameters of the plant are adjusted without any numerical integration. After the adjustment, the adjusted estimators are transported to the control part and are used to control the input generation over an interval, which differs from the interval for estimation since computational time is necessary for estimator adjustment. We show analytically and numerically that the computational burden of the proposed system is less than the usual hybrid adaptive system and that the proposed adaptive control system is globally stable in spite of including a time delay for the estimator adjustment in the control system.