Abstract: Multiple fixed models and adaptive models are set up to cover the uncertainty of a SISO discrete time system with unknown and jumping parameters, and a multi-model adaptive pole-assignment controller is proposed by using a switching index function. On the premise that closed-loop stability is guaranteed, this kind of multi-model adaptive control (MMAC) can improve the transient response greatly. At the same time, localization method is introduced to reduce the computation burden of multi-model algorithm without loss of accuracy.