Abstract In this study, laboratory tests using six-cycles of nitrogen (N2) injection were carried out to investigate nitrogen flushing effect on coal seam gas recovery. In each test, N2 was injected for 200 min into a coal sample saturated with carbon dioxide (CO2), in a triaxial-loaded cell. Nitrogen injection was paused for about 20 h to allow gas desorption from the coal sample before commencing the next cycle. It was observed that CO2 levels of the outlet flow recovered from 3%-7% to 20%–40%, with no significant change in gas flow rate. The displaced CO2 in each cycle accounts for 4–7% of the initial total gas volume. Analysis of the test results indicated that at least 22 injection cycles were required to drop the residual gas content below the threshold limit value (TLV). The test results also confirmed that N2 injection can promote gas desorption from coal matrix and reduce un-desorbable residual gas content. In comparison with the continuous injection method, the flushing efficiency was lower for the cyclic injection method (5.6 days vs 21.5 days). However N2 consumption was about 60% less, and its utilization ratio was higher for cyclic injection method. Additionally, a post-injection effect of seam gas desorption was observed following each injection cycle. This study demonstrated that cyclic N2 injection method has several potential advantages for different field applications. This injection method can improve coal seam gas recovery with much reduced N2 consumption, and in particular, reduce the cost for separating gas mixture product in nitrogen enhanced coalbed methane recovery (N2-ECBM) project.