This study presents a high‐precision Cd isotope measurement method for soil and rock reference materials using MC‐ICP‐MS with double spike correction. The effects of molecular interferences (e.g., 109Ag1H+, 94Zr16O+, 94Mo16O+ and 70Zn40Ar+) and isobaric interferences (e.g., Pd, In and Sn) to Cd isotope measurements were quantitatively evaluated. When the measured solution has Ag/Cd ≤ 5, Zn/Cd ≤ 0.02, Mo/Cd ≤ 0.4, Zr/Cd ≤ 0.001, Pd/Cd ≤ 5 × 10−5 and In/Cd ≤ 10−3, the measured Cd isotope data were not significantly affected. The intermediate measurement precision of pure Cd solutions (BAM I012 Cd, Münster Cd and AAS Cd) was better than ± 0.05‰ (2s) for δ114/110Cd. The δ114/110Cd values of soil reference materials (NIST SRM 2709, 2709a, 2710, 2710a, 2711, 2711a and GSS‐1) relative to NIST SRM 3108 were in the range of −0.251 to 0.632‰, the δ114/110Cd values of rock reference materials (BCR‐2, BIR‐1, BHVO‐2, W‐2, AGV‐2, GSP‐2 and COQ‐1) varied from −0.196‰ to 0.098‰, and that of the manganese nodule (NOD‐P‐1) was 0.163 ± 0.040‰ (2s, n = 8). The large variation in Cd isotopes in soils and igneous rocks indicates that they can be more widely used to study magmatic and supergene processes.