Since migration causes socio-economic and population changes in urban area, it is considered a significant factor in forecasting demands in a number of urban socio-economic categories. In spite of its importance in planning, however, the existing model lacks in comprehensive explanation of the origins (from where) of migration, the destinations (to where), whom, why, and by how much volume migration takes place inter-regionally. This study is designed to develop a mirgration model for analyzing and forecasting inter-regional migrations, concerning selectivity and determinants of migration which are affected by spatial heterogeneity. In this research, migration model is classified into three levels: Level 1) gender and age-group inter-immigration locations divided by destination; Level 2) by different migration location within an area; and Level 3) by city and gun regional level. Analyzing this model, a hierarchical generalized linear model is employed with consideration on Poisson distribution. Due to spatial effects in level 3, these variables have space-specific effects whose influence become high or relatively low in a specific location. Finally, compared to the existing model, this model’s forecasting quality of gender and age-group migration was precise enough to be used in the real-world urban and regional planning.