Abstract: Under the background of industrial big data and intelligent manufacturing, data-driven intelligent fault diagnosis technology has been regarded as the key enabling technology to ensure the safe and reliable service of high-end equipment, and to realize equipment health management and intelligent operation and maintenance. Aiming at the problem that the existing intelligent fault diagnosis models cannot comprehensively yield the integrated advantages of superior diagnostic accuracy, strong anti-noise robustness, high computational efficiency and strong hyper-parameter robustness, this paper proposes a novel spectral ensemble sparse representation classification model-driven super-robust intelligent diagnostic method. The proposed method firstly designs a vibration data augmentation strategy based on cascade segmentation operators to improve the quantity and quality of vibration data samples. Then, the proposed method utilizes the spectral features of vibration signals to realize the dictionary atom design and constructs a spectral ensemble dictionary design strategy considering spectral feature fusion to improve the reconstruction ability of spectral sparse representation dictionary. Finally, the proposed method develops an intelligent recognition strategy based on the spectral sparse approximation error minimization-based criterion to achieve intelligent diagnosis of health status of test samples. The proposed method is validated on a planetary gear transmission fault dataset and the results show that the proposed intelligent diagnosis method can realize the integration of superior diagnosis accuracy, strong noise robustness, high computational efficiency, and strong hyperparameter selection robustness, and its diagnosis results are superior to the existing advanced methods, showing an important prospect for data-driven intelligent fault diagnosis of industrial equipment.