Apatite chemical composition has often been invoked for appraising the magmatic copper (Cu) fertility, because trace elements in apatite hold important clues for tracing magma composition, oxidation states, and crystallization processes. However, low-dimensional Cu fertility discriminants developed on apatite trace elements suffer from significant limitations and uncertainties in practice. Here, machine learning (ML) models including random forests and support vector machines were trained using high-dimensional apatite composition dataset (spanning 20 trace elements) for discriminating ore-bearing magmas from ore-barren magmas. The results suggest that the ML models obtained a higher accuracy (96 %) for identifying the given apatites from ore-bearing samples compared to that of traditional discriminant diagrams (56 %). The feature importance analysis suggests that δEu and Sr are the most significant proxy for distinguishing ore-bearing and ore-barren samples when using high-dimensional ML models. In general, apatites from ore-bearing intrusion have higher δEu and Sr concentration, lower Pb concentration, and elevated Sr/Y ratio than ore-barren samples. Specifically, the elevated δEu and Sr concentration indicate a relatively higher oxidation state and water content in parental magmas, which could have promoted sulfate formation and Cu release and transport. Moreover, the elevated Sr/Y observed in apatite from ore-bearing samples imply the adakite-like composition of the ore-productive magmas, while lower Pb concentration suggests strong fluid participation during magmas evolution. The trained ML model was applied to apatites from the Tampakan district of the Philippines, providing new insights on Cu fertility of pre-, syn- and post-ore intrusions. The general applicability of this model demonstrates that ML-based discriminants developed on mineral trace element data provide new powerful tools for appraising the porphyry Cu fertility.