The objective of this study was to test the feasibility of a new method to improve the accuracy in the estimation of sewage components. Adding to the regression of sewage components with UV (ultraviolet) absorbance values, a proposed method considered an unclear but existing relationship among characteristic of sewage production. Sewage production showed very defined profiles due to the daily human activities. So the main idea was the combination of measuring the UV absorbance values and analyzing the characteristics of the sewage production. For this purpose, 446 sewage samples taken at every 2-h interval for 51 days at a wastewater treatment plant were statistically analyzed using neural network (NN). NN was trained with 350 data sets (about 29 days) of UV absorbance values, flow rate and time. And as a result, it could predict 96 data (12 days) as a validation, indicating that estimation accuracies were improved to higher level than those of the linear regressions. The proposed method could estimate concentrations of total nitrogen (TN) and total phosphate (TP) within practical accuracies as well as total suspended solid.