Bridge inspection and service performance evaluation are critical technologies for ensuring the safe operation of bridges. Utilising a multi-scale perspective, the paper systematically reviews the academic progress and future trends in the field of bridge damage detection and assessment. The research is explored from three different scales: macro, meso, and sub-micro. A thorough analysis of the evolution of bridge feature detection methods is presented at the macro level, illuminating the trend of transformation toward rapid detection technologies based on vehicle responses. At the meso- and sub-micro scales, the complexity of bridge surface damage has resulted in research focusing on recognition methods based on computer vision. In terms of service performance evaluation, the extant methods for short-term bridge condition assessment and long-term condition prediction are summarized. The comprehensive analysis shows that the current bridge inspection technology has been effective in identifying bridge damage features. However, future research should still focus on two directions: macro damage identification based on vehicle response and meso- and sub-micro damage identification based on computer vision. Both directions have shown great application potential. Future research should further optimize vehicle-bridge coupled response models and improve their applicability to different forms of macro-scale damage; Study the mapping relationships between meso- and sub-micro-scale damage images and bridge mechanical characteristics; conduct research on multi-scale damage correlation to improve detection accuracy; and explore more practical evaluation methods for bridge service performance based on engineering practice.