In order to improve accuracy and timeliness of chemical oxygen demand（ COD）

using algorithm which combines ultraviolet absorption spectroscopy method with partial least squares regression to predict COD in water sample.At the same time

influence of turbidity on absorbency of modeling is considered

and the effect on turbidity is compensated.Experimental results show that the proposed method is suitable for different types of sewage

the average relative error is less than 5 %

and the prediction precision is prior to the partial least squares regression model without turbidity compensation.