Abstract: Traditional rock tunnel blasting design methods mainly rely on expert experience and manual drawing. Due to the lack of efficient tools and systems, these methods are characterized by low design efficiency, high costs, and poor adaptability to complex geological conditions, making it difficult to meet modern construction requirements. This study proposes an intelligent rock tunnel blasting design system based on deep learning. The background of this study is the excavation of a three-core arch working face in a horizontal roadway 280 m underground in Beishan underground Laboratory of China. The system input basic blasting requirement parameters, tunnel cross-sectional width, height, and advance per cycle. Through a deep learning model, the system optimizes blasting parameters with deep-learning models, smartly generate borehole layout diagrams in line with construction regulations, and achieve high-efficiency and high-precision blasting design. The core module of the system combines random forest and reinforcement learning, and the stacking ensemble learning method is employed to improve the model’s prediction accuracy and generalization ability, outputting optimized parameters such as the number of boreholes, spacing, and charge amounts. The system is developed using the Python programming language and Django framework, integrating drawing tools to achieve intelligent blasting parameter design, automated borehole layout generation, and visualization. The system is designed with modularity and scalability in mind, covering functionalities such as data processing, deep learning model training, parameter optimization, diagram generation, and blasting log recording. Efficient data interaction and functional integration between modules are achieved through API interfaces. Experimental results show that the system significantly improves blasting design efficiency, increasing design efficiency by more than 30%, enhancing borehole utilization by over 10%, and controlling over-excavation to ≤10 cm. The system significantly reduces reliance on expert experience, improves the accuracy and reliability of blasting design, and offers effective technical support for the field of intelligent blasting.