We investigate the effects of childhood work on migration decisions and patterns later in life using a novel prospective panel dataset from rural Ethiopia. The data were generated through a follow-up tracking survey of 4-14-year-old children at baseline (1999/2000) after sixteen years in 2015/2016. We find that village out-migration was by and large dominated by females and schoolchildren. Compared to schoolingonly, fulltime childhood work significantly reduces the probability of village out-migration later in life. In contrast, those who combined work and study at baseline were highly likely to engage in economic or employment out-migrations. Thus, we presented new evidence in the related literature that besides the existing rural to urban labor migration explanations, childhood conditions in a rapidly changing developing economy setting may also affect children’s long-term migration decisions. The findings also suggest that elimination of full-time child labor should be a long-term human capital policy priority. However, excluding the worst forms of child labor, an attempt for child labor elimination in all its forms could be un(counter)productive. More importantly, rural child education seems to be as critical as enabling the future farm labor to shift from farm to non-farm activities and facilitate the structural transformation process.