Investigation of spatial distribution of oil and gas resource and accurate prediction of the geographic location of its undiscovered resource is signifi cant for reducing exploration risk and improving exploration benefi t. A new method for predicting spatial distribution of oil resource is discussed in this paper. It consists of prediction of risk probability in petroleum exploration and simulation of hydrocarbon abundance. Exploration risk probability is predicted by multivariate statistics, fuzzy mathematics and information processing techniques. A spatial attribute database for sample wells was set up and the Mahalanobis distance and Fuzzy value of given samples were obtained. Then, the Bayesian formula was used to calculate the hydrocarbon-bearing probability at the area of exploration wells. Finally, a hydrocarbon probability template is formed and used to forecast the probability of the unknown area. The hydrocarbon abundance is simulated based on Fourier integrals, frequency spectrum synthesis and fractal theory. Firstly, the fast Fourier transformation (FFT) is used to transform the known hydrocarbon abundance from the spatial domain to the frequency domain, then, frequency spectrum synthesis is used to produce the fractal frequency spectrum, and FFT is applied to get the phase information of hydrocarbonbearing probability. Finally, the frequency spectrum simulation is used to calculate the renewed hydrocarbon abundance in the play. This method is used to predict the abundance and possible locations of the undiscovered petroleum accumulations in the Nanpu Sag of the Bohai Bay Basin, China. The prediction results for the wellexplored onshore area of the northern Nanpu Sag agree well with the actual situations. For the lessexplored offshore areas in the southern Nanpu Sag, the prediction results suggest high hydrocarbon abundance in Nanpu-1 and Nanpu-2, providing a useful guiding for future exploration.