A suboptimal approach to the fixedlag smoothing problem was proposed based on Markov switching system.By introducing stateaugmented system in the interacting multiple model frameworkwhich combined the current state with the past onethe state prediction was made synchronously with the smoothing for the original system.The fixedlag mode probabilities were estimated by using the current observation datawhich is a simple but effective method.The proposed smoothing algorithm was compared with that of the normal IMMEKF filtering algorithm through the trajectory of a maneuvering target.The result shows that the performance of the fixedlag IMM smoothing algorithm is great better than that of laterand the performance tends to get better with the increasing of the lag.