Overnight waste oil has low nutritional value, poor hygiene, and poor edible quality. Illegal traders often mix it with ordinary vegetable oil to reduce production costs and improve profit margins. To protect people's food safety and health, relevant departments must strengthen the crackdown on the adulteration of overnight waste oil. Therefore, this article proposes a one-dimensional convolutional neural network algorithm to assist in the detection method of portable Raman spectroscopy quantitative analysis of overnight waste oil adulterated with sunflower seed oil. Quantitative analysis of simulated adulteration was conducted by preparing adulterated oil products with a wide range and uniform gradient variation. Raman spectral data of mixed oil products with different adulteration concentrations were collected using a portable 633 nm Raman spectrometer. The original spectral data was then preprocessed with baseline correction, noise reduction, and normalization. Finally, the training and testing sets were divided into a 4∶1 ratio, and the model was validated by leave-one-out cross-validation. The results indicate that portable Raman spectroscopy can extract spectral information of two different oils, and the spectral differences between the two oils are mainly concentrated in the two Raman spectral fingerprint regions of 450~2000 cm^-1 and 2500~3100 cm^-1. A quantitative analysis model for 11 types of adulterated oil products was established based on a one-dimensional convolutional neural network algorithm, achieving ideal quantitative analysis. The determination coefficient of the one-dimensional convolutional neural network model test set was 0.9922, and the root mean square error was 0.0279. In summary, the method proposed in this article can achieve a quantitative analysis of common vegetable oil adulteration, and this detection method provides a specific reference value for frontline applications and on-site non-destructive detection.