Current methods to detect and monitor oil contamination in soils rely upon geophysical techniques or direct sampling and testing and their use amongst the global community is limited by their complexity and cost. Recent studies demonstrate that Visible-Near Infrared Spectroscopy (VNIRS) may be a rapid and reliable method but there are limitations due to uncertainties over the effects of soil composition, moisture content and oil degradation. A laboratory-based study has been conducted that examined the VNIRS properties of artificially-prepared and natural soil samples representing different soil compositions and different moisture contents. Synthetic crude oil was added to samples to obtain oil concentrations of 0.25 % to 40 % and results were used to derive two new oil indices termed the ‘Oil Index’ (Oix) and ‘TPH Index’ (TPHix). Analyses of Oix versus % Oil yielded a series of polynomial regressions with a discrete inflection termed the ‘Critical Point’ (CP). This was found to vary with soil composition, and in particular, clay mineralogy. CP also varied with moisture content. Comparison with the literature indicates that the variability described by the Oix-CP was related to sorption/desorption phenomena and may a have direct relation with the sorption capacity of a soil. Analyses of TPHix demonstrate that this more complex index provides a robust method to estimate oil content in soils, especially at low concentrations, though measures will still need to be modified with respect to soil class. Analyses of time-series data demonstrate that VNIRS methods were comparable to conventional gravimetric methods but yield other useful information related to oil degradation. The research, the first of its kind, demonstrates that VNIRS can be used to accurately measure and monitor crude oil contents in soils. Further, the technique provides insight into soil-oil-water interaction where there are few established techniques available to researchers.