The optimum conditions for measuring cadmium content of less than 0.2ppm by flame atomic absorption spectrophotometry were investigated. The cadmium in urine was extracted by APDC-MIBK for the analysis by atomic absorption spectrophotometry after ashing them by a wet method. 1. Optimum conditions by APDC-MIBK and DDTC-MIBK extractions. The acidic aqueous solution was prepared with appropriate amount of 0.IN nitric acid, 5ml of 25% (W/V) sodium potasstum tartarate, 10ml of saturated ammonium sulfate, and 2ml of 2% APDC(or 1 ml of 5% DDTC) chelating agent. The total volume of solution was adjusted to 55 ml and pH to $2{sim}10$ (or$7{sim}10$). The aqueous solution was extracted with 10ml MIBK. Concentration of Triton X-100 did not effect the absorbance for APDC-MIBK extraction of cadmium, but absorbance decreased as the concentration increased for DDTC-MIBK extraction. The sensitivity and detection limits for the cadmium determination from APDC-MIBK extraction were 0.0038ppm and 0.0102, 0.0022ppm and 0.0116 for DDTC-MIBK, and 0.0132ppm and 0.0034 for 0.1N nitric acid. APDC-MIBK and DDTC-MIBK extractions were 3 times higher than 0.1N nitric acid for the sensitivity. 2. Excretion of cadmium in 24-hour urine by APDC-MIBK extraction. Determination of cadmium in urine by atomic absorption spectrophotometry of A.A. (Cd=2 mA) mode and B.C. (Cd=4 mA) mode and B.C. (Cd=4mA, $D_2=20mA$) mode showed some difference (p<0.05). The difference of cadmium determination and recovery according to method of standard additions and standard calibration curve method in urine was not significant (p>0.05, $93.48{pm}11.78%,;94.83{pm}22.00%$). Excretion of cadmium in 24-hour urine collection from normal person and variance analysis within measurement variation was not significant (p>0.05), but between interindividual was significant (0.05). Determination of cadmium content by two different methods of flame atomic absorption spectrophotometry and dithizone colorimetry showed that the results from the two methods can be described by a regression line with a good correlation (y=1.0153x-0.2927, x=Cd by D.C., y=Cd by A.A.S., $r=0.8651^*$, p<0.01).