Hyper-spectral imaging (HSI) is an emerging, hybrid process analytical technology, combining imaging and spectroscopic techniques for food quality monitoring and assessment. While this technique has recently proved popular for food quality assessment in the fruit and seafood industries, there are only a few reported applications of HSI in the dairy industry. The interest in HSI is due to its ability to process a considerable amount of spectral data over a spatial dimension. In this work we analysed three plants all making a specific valuable milk powder. However the milk powder produced by each plant is different and each plant has different key equipment types such as the dryer. It is hypothesised that there is a causal relationship here. In this paper, the potential application of HSI to discriminate between the milk powders produced at the three different plants is presented, specifically with respect to the prediction and monitoring of functional properties such as dispersibility and solubility. Principal component analysis (PCA) was applied on hyper-spectral data extracted from milk powder samples from the three plants. The results showed that the major discrimination between milk powders produced by the different factories occur in principal components (PC) 2 and 3, and not in the first PC as this component correlates to milk powder morphology. Furthermore, the potential of the HSI technique to classify the powder as either on or off-spec at close to real time speeds is explored. The current limitations of this process analytical technique and potential future developments involving HSI in the dairy industry are also discussed.