The development of energetic materials faces many challenges， and the traditional trial-and-error research model often results in long development cycles and low efficiency. With the advancement of data science and artificial intelligence （AI） technologies， a data-driven research model has emerged as a new path for the development of energetic materials. Polycyclic energetic compounds are currently a hot topic in the field of energetic materials， among which nitrogen-containing polycyclic frameworks， due to the presence of π electrons for delocalized resonance and multiple modifiable sites， exhibit enhanced molecular structural stability. At the same time， the presence of energy groups ensures the energy level of the molecules， achieving a good balance between energy and stability， overcoming the inherent contradiction between them. This study briefly introduces the workflow of data-driven development of novel energetic materials， outlines the latest research progress of data-driven methods for the development of nitrogen-containing polycyclic energetic compounds， and finally proposes prospects for the application of data-driven methods in the development of novel energetic materials. Future directions should consider supplementing data volume through means such as data augmentation and governance to improve the accuracy and generalization ability of model predictions. Machine learning models can be used to predict the molecular synthetic feasibility by establishing chemical reaction conditions and synthetic pathways， thereby accelerating the development of novel nitrogen-containing polycyclic energetic compounds.