Aiming at the problem of condition identification and dynamic regulating of running condition domains of a car engine，this paper presents a method to dynamically adjust the running condition domains on the basis of speed and load forecasting. The method can well combine the advantages of adaptive neuro fuzzy inference for speed and load prediction and immune feedback regulating for running condition domains. The relation between the operating condition domains and the period and amplitude of the speed and load deviation is discussed. Simulation results show that the method can dynamically adjust the running condition domains and improve the adaptability of an engine under complex pavement.