Traditional acoustic design of heating, ventilation and air conditioning (HVAC) system mostly relies on the experience of designers, with significant repetition and minimal automation. To achieve the goal of intelligent design, methods such as directed acyclic graph, weighted directed adjacency matrix, and visualization were applied in the acoustic design of ventilation duct system. An algorithm was proposed to automatically calculate the noise levels at any place, whether in the duct or a specific room. The key components such as building 3D models, designing noise attenuation diagram, determining acoustic path, solving acoustic unit, and visualizing acoustic results were included in this system. The results show that this system is robust, accurate and quick to respond. The deviation between results from traditional methods and the proposed algorithm is no more than ± 2%, and the computation time is less than 0. 2 s, whereas the running time for some small and medium ventilation systems is within 0. 1 s. The results based on this algorithm can be combined with 3D models to visualize the change of noise values and provide a reference in the design process of HVAC systems.