Protein interaction networks underlie countless biological mechanisms. However, most protein interaction predictions are based on interspecies biological evidence (knowledge-based predictions) that are biased to well-known protein interaction or physical evidence that exhibits low accuracy for weak interactions and requires high computational power. In this study, a novel method has been suggested to predict protein interactions using interaction energy distribution. Protein interactions with specific partners can be predicted by investigating narrow funnel-like interaction energy distribution, exhibiting a stable state in a specific orientation. In this study, it was demonstrated that various protein interactions with specific interaction partner, such as protein interaction with kinase and E3 ubiquitin ligases, have narrow funnel-like interaction energy distribution. Moreover, algorithm and deep learning model for prediction of general protein interaction partner and substrate of kinase and E3 ubiquitin ligase were developed. The prediction accuracy of this method is similar to or even better than that of yeast two-hybrid screening, which is the most widely applied interaction screening method. Ultimately, this knowledge-free protein interaction prediction method will broaden our understanding of protein interaction networks.