With the development of modern weapon platform, the position of multi-sensor management is becoming more and more important in the information fusion system. In the air defense warfare under information-based condition, establishing Air-Ground air defense situation awareness network is a technical assurance to capture informational preponderance, meanwhile realizing scientific and reasonable allocation of sensors resource is an important method to improve the gathering efficiency of air defense network. We have to manage the limited sensors resource scientifically and reasonably so as to improve air defense warfare effectiveness. The optimization model based on information gain and target threat degree is proposed for multi-sensor management. This model, through calculating rough entropy of the sensor to the goal, obtains the greatest information content of each sensor to each goal. By taking this as a price function and in view of target threat degree, the optimized assignment of the multi-sensors to the multi-goals is done by using linear programming, the simulation result indicates that this model is effective and reasonable.