In recent years, palm oil has been one of the most consumed in the world due to its physicochemical characteristics. It is widely used in the food industry, outweighing the consumption of soybean oil. The specific physical-chemical characteristics of each oil are related to its purity, processing, climatic conditions beyond the maturation stage of the fruit. The use of nondestructive tools, such as NIR spectroscopy, combined with mathematical and statistical methods. To classify samples and predict quality parameters is promising. The objective of this work is the application of a low cost NIR instrument to classify palm oil according to the heat treatment used and to obtain a multivariate calibration model to predict the acidity content. Spectral information of samples was obtained in the NIR range (900-1700 nm) . Principal Components Analyses presented spectral difference between two different thermal treatments. The heat treatment is used to inactivate the enzymes. When the sterilization conditions are not performed correctly, the enzyme is not inactivated and can degrade the oil by changing its quality. The LDA classification method was applied to selected wavelengths, obtaining 95% accuracy in the classification of samples, other methods of classification were k-NN and PLS-DA. These methods presented a high classification index with sensitivity and selectivity above 97%. To predict the acid content of the samples, PLS calibration model was proposed, obtaining a correlation coefficient of 91%. The quality of the oils is directly related to acidity, since this is an important parameter in the identification of degradation. These results demonstrate that it is possible to use a low cost NIR tool to characterize palm oil according to quality parameters.