The axle-box bearing is a crucial component of rail vehicles, and its service condition has a direct impact on the safety and stability of the entire vehicle system. Envelope spectrum analysis based on spectral coherence (SCoh) is an effective tool for revealing the underlying periodic mechanism caused by bearing defects. The improved envelope spectrum (IES) via Candidate Fault Frequencies Optimization-gram (IESCFFOgram) provides an efficient approach for identifying frequency bands without relying on sparsity indicators or fault characteristic frequency. However, the applicability of the IESCFFOgram is limited to the analysis of vibration signals with a single frequency band. Therefore, a new method referred to as combined IES (CIES) is proposed based on candidate fault frequencies to generate a spectrum tool carrying diagnostic information dispersed in multiple frequency bands. The proposed method is tested and validated using axle-box bearing datasets from a test rig. Comparisons with state-of-the-art methods demonstrate the superiority of the proposed CIES method in identifying multiple informative frequency bands under complex operating scenarios.