The selection of an appropriate method for decomposition of anomaly patterns of stream sediment geochemical data is a challenging issue, as the geochemical anomalies of stream sediment data exhibit complex anomalous patterns. This study compares the results of delineation of multi-element geochemical anomalies of stream sediment data, derived via number–size (N-S) and concentration–area (C-A) fractal models and U-spatial statistics. Principle component analysis (PCA) was applied for Cu, Mo, and Au geochemical signatures to derive a multi-element geochemical signature associated with porphyry Cu deposits. Different geochemical populations of the multi-element geochemical signature were delineated by three modeling methods. We used the location of known porphyry Cu occurrences and employed two criteria of the Student’s t value and the normalized density index for the comparison of modeling methods. Results revealed the superiority of the U-spatial statistics method over the rest of models. Furthermore, the results of C-A model were better compared with those of N-S model.