Mountain farming communities in Pakistan are exhibiting an increased rate of rural-to-urban migration and a rapid growth in the non-farm sector, which has threatened the sustenance of agricultural activity. This study examined the determinants of farm succession using a logit regression model and employed a multinomial logit regression model to study the factors influencing the future occupational choices of the potential farm successors. The study was based on quantitative survey data obtained from 421 farm managers and 155 potential farm successors and qualitative data from 12 key informants from two different districts in Gilgit-Baltistan. The survey results show that around 67% of the farmers had a potential successor. Farm succession was mainly explained by farmer characteristics (e.g., farmer age, gender and education), farm characteristics (e.g., farm size, specialization in horticulture, etc.) and agricultural income. Regarding the occupational choices, part-time farming (66%) was the most commonly reported choice. The results indicate that it was mainly farm successors’ personal characteristics (such as age, education and marital status) and agricultural income that led to the choices “undecided” and “exit”, whilst farm characteristics (e.g., farm size) and the main farm operators’ non-farm activity were significantly associated with the choice “part-time”. Policies aimed at improving the local income situation and investments in skill-building and infrastructure development can assist in farm sustenance.