OBJECTIVE： To establish a determination method for the concentration of cajanonic acid A （CAA） in liver microsome incubation system， and to compare the metabolism characteristics of it in different species of liver microsomes. METHODS： CAA was dissolved in liver microsome incubation system of rat， Beagle dog and human initiated by reduced nicotinamide adenine dinucleotide phosphate （NADPH）， and was incubated in water at 37 ℃. The reaction was terminated with acetonitrile at 0， 5， 10， 15， 30， 45 and 60 min， respectively. Using genistein as internal standard， the concentration of CAA in  different incubation systems was determined by UPLC-MS/MS. The determination was performed on Waters BEH C18 column with mobile phase consisted of water （containing 0.1％ formic acid）-acetonitrile （containing 0.1％ formic acid） （45 ∶ 55， V/V） at the flow rate of 0.25 mL/min. The column temperature was 30 ℃， and the sample size was 2 μL. The electrospray ionization source was used to the select reaction monitoring mode for negative ion scanning. The ion pairs for quantitative analysis were m/z 353.14→309.11 （CAA）， m/z 269.86→224.11 （internal standard） respectively. The residual percentage and enzymatic kinetic parameters of CAA in different incubation systems were calculated according to the mass concentration of CAA at 0 min. RESULTS： The linear range of CAA was 0.05-20 μg/mL； the limit of quanti- tation was 0.05 μg/mL， and the lowest detection limit was 0.01 μg/mL. RSDs of intra-day and inter-day were lower than 10％； relative errors ranged －4.83％-8.94％； extraction method and matrix effect did not affect the determination of the substance to be measured. At 60 min of incubation， residual percentages of CAA in rat， Beagle dog and human liver microsomes were（62.79±9.99）％，（64.07±11.59）％，（96.66±5.71）％， respectively. The half-life period （72.19， 68.61 min） of CAA in rat and Beagle dog liver microsomes were significantly shorter than human liver microsome （364.74 min）. The clearance rates [0.019 2， 0.020 2 mL/（min·mg）] were significantly higher than human liver microsome [0.003 8 mL/（min·mg）] （P＜0.05）. CONCLUSIONS： Established UPLC-MS/MS method is simple， rapid， specific and sensitive， and can be used for the determination of CAA concentration in liver microsome incubation system and the study of metabolism stability in vitro. The stability of CAA metabolism in rat and Beagle dog liver microsomes are poorer than human liver microsome.