From four aspects including weak form efficiency test、semi-strong form efficiency test、no-arbitrage equilibrium test and market anomaly test, this paper researched the information efficiency of China government bond market. The empirical results shows that the information efficiency of China bond market is low. The market does not achieve complete weak form efficiency , and does not achieve semi-strong efficiency as well.On the weak form efficiency test, this paper studied monthly return series of different maturity portfolios using autocorrelation test、run test and variance ration test.In light of the presence of significant autocorrelation in some series , we can reject Random Walk Hypothesis .On the semi-strong efficiency test , this paper used regression method and event study method. Results tell us that macro economic variables can not forecast government bond month returns ; results also tell us that government bond price can react to macro announcement rapidly, but there are long lags in this reaction, which means government bond market has not achieve semi-strong efficiency.No-arbitrage equilibrium is the necessary result in an efficient market. But the empirical results show there are significant arbitrage opportunities between the inter-bank market and exchange market. As a complement to the weak form efficiency test , this paper test theWeekend effectin government bond market. The results show the presence of significantWeekend effect, but this anomaly can be interpreted bySettlement Regime Hypothesis, so theWeekend effectcan not be used to reject the weak form efficiency hypothesis.