The spatial distribution of a basin hydrological network is usually unreasonable, and the network may contain a large number of redundant stations and monitoring blank areas. In this regard, the 4-year and mean daily runoff data from 27 sites of the Lower Illinois River Basin is used in this study to achieve an optimization of hydrological network. Three objective functions, which are joint entropy ratio, redundancy ratio and Nash-Sutcliffe efficiency coefficient(NSE), are constructed to quantitatively evaluate the information of station network, based on three criteria of maximum joint entropy, minimum total correlation and minimum data fluctuation. The results show that a preliminary optimized site combination composed of 11stations can provide the same amount of information as the original station network, and at the same time, the redundant information can be reduced by 74. 2%. The total amount of information and redundant information were increased by 37. 1% and 22. 3% respectively by adding 3 sites on the basis of preliminary optimized site combination. Compared with the original station network, the number of optimized sites is 14, the total amount of information is increased by 37. 1%, and the amount of redundant information is reduced by 68. 5%.