Alluvial small-scale gold mining (ASGM) mining in the Amazon is expanding fiercely, generating severe environmental degradation including the disappearance of primary forests. Different factors motivate the expansion of mining activities. However, their interaction, influence and new directions when expanding to new gold mining areas are not well understood. Anticipating this growth is important to safeguard protected areas or to implement strategies to mitigate the related social and environmental impacts. This study uses machine learning (ML) techniques to explore gold mining expansion in the Peruvian Amazon, and to identify possible future hotspots. Unsupervised learning algorithm and a classification model (ie, random forest) are used to achieve this objective. Results show that ML techniques allow anticipating new mining areas and subsequent deforestation. Two important areas of high mining probability were identified: Puerto Pariamanu and Puerto Maldonado. The model demonstrates to be useful to enhance decision making when implementing geo-spatial policies.