The thermal maturity of oils extracted from inclusions and the fluorescence colours of oil-bearing fluid inclusions have been measured in 36 sandstone samples from Australasian oil fields. The inclusion oils were analysed using an offline crushing technique followed by GC-MS. A maturity assessment was made for each inclusion oil using 25 molecular maturity ratios, including a newly defined dimethyldibenzothiophene ratio (DMDR). Each inclusion oil was placed in one of 4 maturity brackets, approximately equivalent to early, mid, peak and post oil generation windows. The fluorescence colours of oil inclusions were visually-discriminated into "blue, "white" and "yellow plus orange" and their proportions estimated using point counting techniques. Sixteen samples have > 85% of oil inclusions with blue fluorescence, whilst other samples have more variable fluorescence colours. One sample has 100% of oil inclusions with yellow plus orange fluorescence. The results show that samples containing mainly blue-fluorescing oil inclusions have thermal maturities anywhere within the oil window. In particular, the molecular geochemical data strongly suggests that oil inclusions with blue fluorescence can have relatively low maturities (calculated reflectance < 0.65%), contrary to the widely applied assumption that blue fluorescence colours indicate high maturities. Samples containing mainly white-fluorescing oil inclusions have maturities anywhere within the oil window and cannot be distinguished using molecular geochemical parameters from samples containing mainly blue-fluorescing oil inclusions. Though few in number, samples with mainly yellow and orange-fluorescing oil inclusions tend to have maturities in the lower half of the oil window. The data presented strongly suggest that although the relationship between API gravity and the fluorescence properties of crude oils is well established, the extension of this relationship to the use of the fluorescence colours of oil inclusions as a qualitative thermal maturity guide is not justified. Fluorescence colour depends in the first instance on chemical composition, which is controlled not only by maturity but by several other processes. For example, inclusions in samples from below current or residual oil zones in the Timor Sea contain a high proportion of yellow- and orange-fluorescing oil inclusions compared to the overlying oil zones, which are dominated by blue-fluorescing oil inclusions. This observation is interpreted to be due to water washing causing molecular and gross fractionation of oils prior to trapping. Fractionation of the gross composition of oil during the inclusion trapping process may also be a significant controlling process on the fluorescence colours of oil inclusions, due to the preferential adsorption of polar compounds onto charged mineral surfaces. A trapping control is strongly supported by synthetic oil inclusion work. Care should be taken when interpreting the charge history of samples containing oil inclusions with mixed fluorescence colour populations, such as those from the Iagifu-7x well in the Papuan Basin. It is possible that the different colour populations represent a single oil charge, with oil inclusions trapped under slightly different conditions or at slightly different grain surfaces, rather than multiple migration events.