Objective: To establish a continuous flow injection analysis with potassium permanganate for determination of oxygen consumption in water. Methods: The water samples and acid potassium permanganate working solutions were mixed using the continuous flow bubble spacing, and subjected to online heating reaction at 97 ℃. The peak height of the electrical signal of potassium permanganate was measured at the maximum absorption wavelength of 520 nm, and the standard substance, sulfuric acid concentration and potassium permanganate concentration were optimized according to the peak height of the electrical signal. The standard curve was plotted to measure the limit of detection, the limit of quantification, and spiked recovery rate of the method. The CODMn concentration was determined in 40 drinking water samples using acid potassium permanganate titration and continuous flow injection analysis using potassium permanganate, and the determination results of the two methods were compared with paired t-test. Results: Glucose was selected as the standard substance, and the mixture of 17.5% sulfuric acid and 3.2 mmol/L potassium permanganate was selected as the working solution. CODMn had a good linear relationship at concentrations of 0 to 6.00 mg/L, with a correlation coefficient of 0.999 and higher, a detection limit of 0.013 mg/L and a quantitation limit of 0.043 mg/L, respectively. The spiked recovery rates were 90.00% to 105.00% in 40 drinking water samples, with relative standard deviations of 0.12% to 1.36%. The determination results of two permanganate index standard substances were all within the range of standard values. The relative errors of CODMn concentration were 1.55% to 9.26% between the continuous flow injection analysis using potassium permanganate and acid potassium permanganate titration, and there was no significant difference (t=2.023, P=0.185). Conclusion The established continuous flow injection analysis with potassium permanganate is feasible for batch determination of oxygen consumption in water.