In this paper we discuss how to generate the command value for the optimal regulator in an automotive antiskid system. First, the behavior of the vehicle at braking is expressed as a mathematical model with the formulation by physical consideration and identification of the hydraulic system by statistical methods. An optimal regulator with additional integral is applied to the automotive antiskid control in order to make the each wheel speed follow any command value. However, the desired command value to stop the vehicle efficiently and stably is dependent on ambiguous road surface conditions. Thus, how to determine the desired command value under the moment-to-moment conditions is most important. A method for inferring the conditions is developed using fuzzy logic, with three fuzzy variables expressing the conditions adequately. On the basis of the inference, the ideal command values are generated. Outstanding control performance and good adaptability are obtained in vehicle experiments. Consequently, the Expert Antiskid System, employing modern control theory and fuzzy logic, can stop a vehicle efficiently and stably under any condition.