The main objective of this work is evaluating the influence of water content and average droplet size upon the near-infrared (NIR) spectra collected during water-in-crude oil emulsions synthesis and observing whether NIR spectroscopy may be used for predicting these properties simultaneously. It is shown that NIR spectra are sensitive to changes of the water content and average droplet size and that standard empirical models [partial least-squares (PLS)] may be built to correlate these properties and total absorbance at the NIR region properly. Finally, it is shown that these models, built with off-line experiments, allow for the online evaluation of average droplet size and water content in water-in-crude oil emulsions with NIR spectroscopy when low water content (<5 wt %) is involved.