Petroleum in the deep- super deep strata has become an important exploration target. In the Gaoshiti- Moxi area of the Sichuan basin, more than one trillion cubic meters of gas reserves have been proven in the deep Sinian and Cambrian. However, many outstanding questions remain unsolved, such as gas- source identification. Solid bitumen has been found widely distributed in the reservoir, leading most geochemists and geologists to believe that the natural gases might have originated from the cracking of crude oil in situ. Thus, an important area of enquiry for deep gas exploration involves looking for paleo- oil accumulations in the Sichuan basin. However, this avenue of research cannot explain an apparent paradox there namely, carbon isotopes of methane, the dominant component of natural gases, are heavier than those of bitumen. Therefore, most geochemists think that organic matter evolves along an isotope fractionation trend under high temperature and high pressure conditions in deep deposits that are very different from those in shallow- middle deposits. In this study, we conducted detailed analysis of the Gaoshiti- Moxi reservoir, and found that there are two stages of bitumen accumulation in the reservoir. The early- stage bitumen is sparsely distributed. It is low in abundance and relatively enriched in 13C (δ13C >-33‰). All this suggests that the early bitumen would be the residual product of crude oil cracking. The late- stage bitumen is abundant and widely distributed in the reservoir. The main filling of petroleum for late bitumen should have been at a burial depth with temperatures over 160~180 ℃, corresponding to the condensate and wet gas window. Their δ13C values arerelatively light (-35. 4‰ on average). Those condensates depleted in δ13C would be generated directly from kerogen cracking rather than from crude oil cracking, because condensate from the cracking of kerogen should be lighter by -3‰ to -5‰ in δ13C than those from the cracking of oil. Thus, all this suggests isotope tracing is still valid in deep deposits. The Sinian natural gas present (δ13CCH4 -34‰~-32‰) ought to have originated primarily from late gas liberated directly by the Qiongzhushi shale. The filling mechanisms would be the fluid pressure difference between the Qiongzhushi shale and the late Cretaceous Sinian strata. The fluid pressure difference causes natural gas to migrate from the upper or lateral Qiongzhushi shale down to the Sinian reservoir. Therefore, the bitumen and the natural gases are not related in origin. The bitumen originated from the cracking of the early filled crude oil and condensate, followed by the natural gases. Accordingly, our gas exploration work in the deep of the Sichuan basin should break from the conventional wisdom of ‘Looking for the paleo- oil accumulation area as the best exploration target for natural gases’. Thus, the top of the inherited paleostructure would not be the primary requirement if we do not think it important to look for paleo- oil accumulation. We should combine all petroleum system factors- including source, reservoir, and seal- as one system, and direct our attention to those carbonate reservoirs, which are well- conserved and are still being fed by the Qiongzhushi Shale. Our study confirms the validity and importance of isotope fractionation in a high- temperature/high- pressure environment, and helps to reveal the carbon cycle and hydrocarbon formation/evolution in the deep deposits.