Previous analysis and modelling of interregional migration in China have treated migrants as a homogenous group. The flow of skilled migration is the focus of recent research. However, the skilled and less‐skilled migrations have not been systematically analysed and compared in terms of their determinants. Previous modelling of interregional migration in China does not take network autocorrelation into consideration. This paper attempts to fill this research gap by modelling skilled and less‐skilled migrations in China using the eigenvector spatially filtered method. It is found that both skilled and less‐skilled migrants tend to move away from the interior to the coastal region. Results from the eigenvector spatially filtered negative binomial regression model show that, compared with the migration of less‐skilled people, the migration of skilled people is less influenced by the friction of distance, the regional unemployment rate, and the concentration of foreign investment but is more affected by the regional wage disparity. With respect to the effect of amenities, climatic amenities exert a strong influence on skilled migration but have positive effect on less‐skilled migration at origin and no effect at destination. Quality medical services are influential for the migration of less‐skilled people to destinations but no effect on skilled people. Copyright © 2016 John Wiley & Sons, Ltd.