This study analyzed the factors affecting the propensity internal migration of Indonesian elderly based on age classification of the elderly, i.e., young elderly, middle elderly, or old elderly. It used multinomial logit model that explored the affect of interaction between internal and external factors on elderly migration. This model was chosen beacuse the dependent variable had polichotomous scale, nominal or ordinal scale with more than two categories. The estimation results indicated the migration propensity of the elderly affected to internal factors, i.e., gender, health, education, employment status, and presence spouse. Meanwhile, external factors, that is, social protection does not affect to the elderly migration. Furthermore, the young women elderly migrants with good health and education background, still working, and lives with their spouse would have a higher propensity to migrate. Instead, these results differed for middle and old elderly migrants. Keywords Elderly Migration; Multinomial Logit; Population Ageing