Abstract: A trading network model is introduced based on the basic rules of limit order book. Through the simulation analysis of the network model, not only the statistical characteristics of degree distribution, clustering coefficient, average path length, degree correlation and so on are described, but also the stability of the network and the relationship between the degree of the network and the trading volume are studied. The feature of small world and scale free is obviously shown. Particularly, the results show the relatively stable clustering coefficient which is not dependent on the network scale, growing average path length with linear slope as the network size increasing logarithmically and the positive correlation between the degree of the network and the trading arrival rate. Meanwhile, the investment network has a high stability for random attacks, while it’s lower under the selective attack.