<p>This paper proposes an algorithm for adjusting the damper mode of a semi-active suspension system according to the driving conditions. This algorithm consists of driving condition recognition and damper mode change. The algorithm's validity is demonstrated through correlation analysis and parametric studies, which show that the optimal damping coefficient for each micro-trip is determined. Driving data are labeled using correlation results and k-means clustering to develop driving condition recognition for damper mode change. The recognition algorithm is constructed based on the analysis of adjustable parameters and has an average accuracy of 79%. Damper mode change is configured based on the driving conditions and clustering results. Unlike conventional controllers that perform instantaneous feedback control, the proposed method determines the damper mode based on long-term driving conditions and maintains it accordingly. Under random road profiles, the proposed algorithm achieves a level of ride comfort comparable to conventional controllers while significantly reducing control effort. Due to its simple structure and use of only in-vehicle sensors, the proposed method presents an efficient and practical solution for implementation in production vehicles.</p>