The trace element composition of igneous and hydrothermal magnetite from 19 well-studied porphyry Cu ± Au ± Mo, Mo, and W-Mo deposits was measured by laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) and then classified by partial least squares-discriminant analysis (PLS-DA) to constrain the factors explaining the relationships between the chemical composition of magnetite and the magmatic affinity and porphyry deposit subtypes. Igneous magnetite can be discriminated by relatively high P, Ti, V, Mn, Zr, Nb, Hf, and Ta contents but low Mg, Si, Co, Ni, Ge, Sb, W, and Pb contents, in contrast to hydrothermal magnetite. Compositional differences between igneous and hydrothermal magnetite are mainly controlled by the temperature, oxygen fugacity, cocrystallized sulfides, and element solubility/mobility that significantly affect the partition coefficients between magnetite and melt/fluids. Binary diagrams based on Ti, V, and Cr contents are not enough to discriminate igneous and hydrothermal magnetite in porphyry deposits. Relatively high Si and Al contents discriminate porphyry W-Mo hydrothermal magnetite, probably reflecting the control by high-Si, highly differentiated, granitic intrusions for this deposit type. Relatively high Mg, Mn, Zr, Nb, Sn, and Hf but low Ti and V contents discriminate porphyry Au-Cu hydrothermal magnetite, most likely resulting from a combination of mafic to intermediate intrusion composition, high chlorine in fluids, relatively high oxygen fugacity, and low-temperature conditions. Igneous or hydrothermal magnetite from Cu-Mo, Cu-Au, and Cu-Mo-Au deposits cannot be discriminated from each other, probably due to similar intermediate to felsic intrusion composition, melt/fluid composition, and conditions such as temperature and oxygen fugacity for the formation of these deposits. The magmatic affinity of porphyritic intrusions exerts some control on the chemical composition of igneous and hydrothermal magnetite in porphyry systems. Igneous and hydrothermal magnetite related to alkaline magma is relatively rich in Mg, Mn, Co, Mo, Sn, and high field strength elements (HFSEs), perhaps due to high concentrations of chlorine and fluorine in magma and exsolved fluids, whereas those related to calc-alkaline magma are relatively rich in Ca but depleted in HFSEs, consistent with the high Ca but low HFSE magma composition. Igneous and hydrothermal magnetite related to high-K calc-alkaline magma is relatively rich in Al, Ti, Sc, and Ta, due to a higher temperature of formation or enrichment of these elements in melt/fluids. Partial least squares-discriminant analysis on hydrothermal magnetite compositions from porphyry Cu, iron oxide copper-gold (IOCG), Kiruna-type iron oxide-apatite (IOA), and skarn deposits around the world identify important discriminant elements for these deposit types. Magnetite from porphyry Cu deposits is characterized by relatively high Ti, V, Zn, and Al contents, whereas that from IOCG deposits can be discriminated from other types of magnetite by its relatively high V, Ni, Ti, and Al contents. IOA magnetite is discriminated by higher V, Ti, and Mg but lower Al contents, whereas skarn magnetite can be separated from magnetite from other deposit types by higher Mn, Mg, Ca, and Zn contents. Decreased Ti and V contents in hydrothermal magnetite from porphyry Cu and IOA, to IOCG, and to skarn deposits may be related to decreasing temperature and increasing oxygen fugacity. The relative depletion of Al in IOA magnetite is due to its low magnetite-silicate melt partition coefficient, immobility of Al in fluids, and earlier, higher-temperature magmatic or magmatic-hydrothermal formation of IOA deposits. The relative enrichment of Ni in IOCG magnetite reflects more mafic magmatic composition and less competition with sulfide, whereas elevated Mn, Mg, Ca, and Zn in skarn magnetite results from enrichment of these elements in fluids via more intensive fluid-carbonate rock interaction.