Spectroscopy is a method of analyzing the physical properties of an object using the difference in light spectrum emitted and absorbed by the material. Various spectroscopic methods including visible light and near-infrared light have been used to evaluate the properties of many targets in various fields by enabling non-destructive and rapid inspection of objects. In the case of spectral analysis of agricultural products, there have also been numerous examples of applications such as evaluation of sugar content of fruit, moisture analysis of wood, and germination rate of seeds. In this study, the same spectroscopic technique was applied and evaluated for two targets: moisture content of leaves and rancidity of edible oil. Water content of litter leaves a great influence on the occurrence of forest fires, and rancidity of edible oil is an index of quality of edible oil. The predictive model for the moisture content of leaves and the rancidity of edible oil was constructed using visible and near infrared spectroscopy for litter and edible oil. The water content of litter and the rancidity of edible oil were measured by the experimental method and the partial least square regression(PLSR) method was used as the predictive model. The rancidity of oil was measured by three indicators: acid value, peroxide value, and TBA. Then new model is selected according to wavelength band selection through the wavelength selection method such as iPLS, VIP, Beta-Bootstrap and Beta-peak in order to remove unnecessary parameters. The leaves were measured for near infrared band of 904 ~ 1707nm and oil for visible-near infrared band of 500-1700nm. For the litter leaves, we applied the four techniques of iPLS, Beta-peak, Beta-bootstrap and VIP, and applied two techniques of Beta-bootstrap and VIP for oil. The predicted model for the moisture content of litters was R ^ 2 in excess of 0.92 in the test set, and the predicted model for the acid value of oil was R ^ 2 in excess of 0.8. The accuracy of the model was increased by the wavelength selection method in the litter leaves, but the accuracy of the model was not increased in the case of oil.