Background and Aims: It is urgent need to develop new-effective and well-tolerable medical treatment for endometriosis, but normal drug development is tedious, costly and time-consuming. Artificial intelligence (AI) is a powerful tool to drive discovery of novel druggable targets. Here we aim to use AI to identify novel therapeutic targets for endometriosis treatment. Method: Targets associated with endometriosis were prioritized by PandaOmics (a well-established AI-driven target identification platform). Expression levels of druggable targets were tested in human endometriotic lesions. Functional characterizations of the targets were performed by in-vitro and in-vivo endometriosis models. Viability, proliferation and apoptosis were examined. Results: Analysis of 11 public endometriosis bulk transcriptomics databases revealed 2 lists of high confidence and novel druggable targets. Tow novel targets ENDO01 and ENDO02 were selected based on ranking in PandaOmics, consistency of significant dysregulated expression across comparisons, as well as literature evidence relevant to disease-driving mechanisms. Differential expressions were confirmed in human endometriosis tissues. siRNA targeting ENDO01 and ENDO02 significantly reduced cell viability and proliferation, as well as enhanced apoptosis in endometriotic cells in-vitro; and significantly decreased lesion volume and weight with reduced proliferative and enhanced apoptosis in-vivo. Conclusion: Two novel targets for endometriosis were identified with AI and validated by in-vitro and in-vivo experiments within a short period. The study demonstrated how AI could speed up the novel target discovery process from years to months. For diseases remaining incurable or with high morbidity, the application of AI in target discovery could be a new therapeutic regimen that benefits patients.