The system deviation according to disruption management were measured from the following three aspects: the customer dissatisfaction, operational cost and processing sequence of jobs. Then, the multi-objective model of disruption management based on the lexicographic multiple goal programming was constructed and an improved ant colony optimization was demonstrated. Finally, the effectiveness of model was validated by a case. The computational result proved that, due to the tradeoff between all parties involved in the single machine scheduling, the proposed model was more practical than existed global rescheduling model.