Abstract: The estimation of ore body resources based on 3D modeling technology and geostatistics has been generally accepted, and the optimal selection of various parameters is the key to reasonable estimation.Based on the systematically analyze of topography, geology, borehole catalogue and testing data of a gold mine in Australia, and according to the transitional relationship between the ore-rock boundaries, this paper proposes to use the mineralized domain estimation theory for resource estimation.Firstly, to establish the drilling database, to conduct basic statistics on the original samples, and to determine the grade boundary value of the mineralized domain.Secondly, according to the grade boundary value of the mineralized domain, to establish the entity model and block model of the mineralized domain, and to extract the sample points in the mineralized domain, to carry out outlier processing and rationality test.Thirdly, to make equal-length combination of sample points after outlier processing, to adjust the search radius of different main axes through geostatistical method, and to optimize the best variogram through cross-validation.Finally, to select the appropriate parameters such as search ellipse sphere, number of samples, et al, and Kriging method is used for resource estimation.The overall verification results show that the estimated grade of the block is nearly identical to the average grade curve of the sample point, indicating that the estimated result is reliable.The mineralized domain model is used for resource estimation, and the results are reasonable and reliable.At the same time, it has certain advantages in industrial index demonstration, ore body characteristic analysis, and metallogenic prediction.