The amount of apple juice consumption keeps increasing as it is the one of the most popular agri-food product in most countries. Current analysis method focused on physiochemical parameters that can be measured in fruit juice sample, such as taste, pulp concentration, acidity, brix level, texture, the level of additives. The primary objectives of this study are developing a model to predict apple juice quality and examining whether the model is suitable to predict the quality of apple juice of different quality level. Overall, MIR spectroscopy results into better PCA model compared to Raman spectroscopy. It can be seen that sample groupings of good and bad quality are achieved by PC1, PC2 and PC3. Prediction of test set using LDA model from training set results into 100% of accuracy for all data measured with Raman and MIR spectroscopy, indicating the suitability of the model to predict the quality of different samples measured at the same time and same instrument.