Liquidity differences between banks can be more significant than both“large scales”and“multiple connections”in measuring systemic risk. Firstly，an approach for testing liquidity differences between commercial banks is proposed from the perspective of interbank networks. The power law distributions of lending (outflow intensity of network) and borrowing (inflow intensity of network) between banks are used for measuring interbank liquidity differences. Empirical results indicate that there were huge liquidity differences among large commercial banks in China between 2012 and 2014. In particular，demands for liquidity of large banks appeared to be most significant in 2014. Secondly，two networks，where the large banks are liquidity-demand and liquidity-supply types respectively，are constructed to examine the influence of liquidity differences on risk contagion. Experiment results indicate that，large banks demanding for liquidity can be much more destructive than banks with similar scales and connections in view of risk contagion. This demonstrates that the liquidity position should be the main criterion for detecting systemic risk by regulators. This paper provides an accurate pressure testing and scenario analysis framework for systemic risk.