The main geomagnetic field models of IGRF1900-2000 are used to study the latitude-dependence of the westward drift in the main field. The results show that the latitude-dependence exists in the magnetic components with different wavelengths (m=1-10). The global-average westward drift rate of the component of m=1 is 0.189°/a with the maximum of 0.295°/a at latitudes 40°-45°. The compo-nent of m=2 has an average drift rate of 0.411°/a with the maximum of 1.305°/a at latitude -60°. As for the compo-nents with further shorter wavelengths, the drift is generally restricted in a limited latitude range, and has many smaller drift rates. This latitude-dependence of westward drift can not be explained by rigid rotation of the earth's core. The results of this note also show that there is a negative disper-sion in the westward drift, namely the components of long wavelengths drift faster than those of short wavelengths. This dispersion feature is not in agreement with Hide's MHD model. It is likely needed t The main geomagnetic field models of IGRF1900-2000 are used to study the latitude-dependence of the westward drift in the main field. The results show that the latitude-dependence exists in the magnetic components with different wavelengths (m = 1-10). The global-average westward drift rate of the component of m = 1 is 0.189 ° / a with the maximum of 0.295 ° / a at latitudes 40 ° -45 °. The compo-nent of m = 2 has an average drift rate of 0.411 ° / a with the maximum of 1.305 ° / a at latitude -60 °. As for the compo-nents with further shorter wavelengths, the drift is generally restricted in a limited latitude range, and has smaller smaller drift rates. This latitude-dependence of westward drift can not be explained by rigid rotation of the earth's core. The results of this note also show that there is a negative disper-sion in the westward drift, of the components of long wavelengths drift faster than those of short of. This dispersion feature is not in agreement with Hide's MHD model. It is likely needed t