The invention provides a method of predicting types and contents of other adulterated oil in tea oil. The method includes the following steps: 1) collecting tea oil, maize oil, sunflower seed oil, and rapeseed oil in different types and brands, and doping the maize oil, sunflower seed oil and rapeseed oil with the tea oil at different ratio, and performing 1H NMR detection to obtain corresponding spectrums, and determining an optimum pretreatment method through a PCA score chart; and 2) further establishing qualitative and quantitative models through orthogonal partial least squares-determination analysis (OPLS-DA) and partial least squares (PLS), verifying the models, and finally identifying the models with a prediction set, when the adulteration quantity of the maize oil, sunflower seed oil and rapeseed oil is more than 5%, determination accuracy is higher than 97%. By means of the NMR data with the partial least squares, the volume fraction of adulterated oil can be quantitatively detected. The method is simple in operation and reliable in technology, and can quickly determine the types and quantities of the adulterated oil in tea oil.