The correct and effective delineation of synthetic geochemical anomaly is the key part of geochemical data processing, and plays a fundamental role in ore prospecting result. Traditional methods delineate synthetic geochemical anomaly based mainly on single element anomaly; nevertheless, many anomalies are caused by stratigraphic high background enrichment. In this paper, the authors use the method of Mahalanobis distance to identify geochemical outliers, and delineates synthetic geochemical anomaly. Because Mahalanobis distance is based on the theory of multivariate normal distribution, and hence is the direct promotion and application of multivariate statistical method. It fully considers three parameters, i.e., element mean content, variance and covariance, its computing process is directly based on samples, and therefore it is also the direct promotion of one-dimensional method having a unique anomaly identification function. The results of geochemical data processing applications with a variety of different scales and different sampling media show that the synthetic geochemical anomaly detected by Mahalanobis distance method has advantages of unique boundary, clear boundary, and reducing the effect of man-made anomaly delineation. In addition, with prominent anomaly intensity and single index, it can be used as an important parameter of anomaly evaluation with high probability of finding orebodies (mineralization). This method therefore can be promoted and applied in practice.