Lacustrine and marine oil shales with Type I and Type I-II kerogen constitute significant petroleum source rocks around the world. Contrary to common belief, such rocks show considerable compositional variability which influences their hydrocarbon generation characteristics. A global set of 23 Ordovician - Miocene freshwater and brackish water lacustrine and marine oil shales has been studied with regard to their organic composition, petroleum potential and generation kinetics. In addition their petroleum generation characteristics have been modelled. The oil shales can be classified as lacosite, torbanite, tasmanite and kukersite. They are thermally immature. Most of the shales contain > 10 wt% TOC and the highest sulphur contents are recorded in the brackish water and marine oil shales. The kerogen is sapropelic and is principally composed of a complex of algal-derived organic matter in the form of: (i) telalginite (Botryococcus-, Prasinophyte- (Tasmanites ?) or Gloeocapsomorpha-type); (ii) lamalginite (laminated, filamentous or network structure derived from Pediastrum- or Tetraedron-type algae, from dinoflagellate/acritarch cysts or from thin-walled Prasinophyte-type algae); (iii) fluorescing amorphous organic matter (AOM) and (iv) liptodetrinite. High atomic H/C ratios reflect the hydrogen-rich Type I and Type I-II kerogen, and Hydrogen Index values generally > 300 mg HC/g TOC and reaching nearly 800 mg HC/g TOC emphasise the oil-prone nature of the oil shales. The kerogen type and source rock quality appear not to be related to age, depositional environment or oil shale type. Therefore, a unique, global activation energy (E(a)) distribution and frequency factor (A) for these source rocks cannot be expected. The differences in kerogen composition result in considerable variations in E(a)-distributions and A-factors. Generation modelling using custom kinetics and the known subsidence history of the Malay-Cho Thu Basin (Gulf of Thailand/South China Sea), combined with established and hypothetical temperature histories, show that the oil shales decompose at different rates during maturation. At a maximum temperature of similar to 120 degrees C reached during burial, only limited kerogen conversion has taken place. However, oil shales characterised by broader E(a)-distributions with low E(a)-values (and a single approximated A-factor) show increased decomposition rates. Where more deeply buried (maximum temperature similar to 150 degrees C), some of the brackish water and marine oil shales have realised the major part of their generation potential, whereas the freshwater oil shales and other brackish water oil shales are only similar to 30-40% converted. At still higher temperatures between similar to 165 degrees C and 180 degrees C all oil shales reach 90% conversion. Most hydrocarbons from these source rocks will be generated within narrow oil windows (similar to 20-80% kerogen conversion). Although the brackish water and marine oil shales appear to decompose faster than the freshwater oil shales, this suggests that with increasing heatflow the influence of kerogen heterogeneity on modelling of hydrocarbon generation declines. It may thus be critical to understand the organic facies of Type I and Type I-II source rocks, particularly in basins with moderate heatflows and restricted burial depths. Measurement of custom kinetics is recommended, if possible, to increase the accuracy of any computed hydrocarbon generation models.