Abstract: PetroChina has just launched two important projects – ^“4th PetroChina Oil and Gas Resource Assessment^” and ^“Study of Global Oil and Gas Resource Assessment and Application^”. It is required to establish a unified assessment method to link the appraisal results to the international convention. To bring the projects smoothly under way, it is necessary to make an analysis of the oil and gas resource assessment methods and compare their application characteristics. This paper makes briefings about the study progress and characteristics of three methods – the simple Pareto model (SP), the shifted truncated Pareto model (STP) and the lognormal discovery process model (LDP). The research method for effective key parameters is proposed, namely using iteration to determine the largest oil reservoir and distribution of oil reservoirs in SP distribution model and using least squares to obtain the median number of oil reservoirs, standard variance and distribution of oil reservoir in the STP distribution model. The paper also describes the application process of the three above-mentioned models, summarizing their applicable scope and comparing the results of those methods in predicting the largest oil reservoir (S), quantity of oil reservoirs (N) and total resources (R). It is concluded that the prediction results of STD distribution model are good when the exploration degree is high with the data meeting the requirements of the model. The SP distribution model is superior when the exploration degree is low with less data made available. The LDP model has a wide application scope and is superior for the area where the exploration degree is at high and middle levels. The SP model can forecast S and R very well but with an extremely large N; The STP model can forecast S very well, but gives large errors on N and R; The LDP model can forecast N and R very well with a worse estimate of S.