Distributed fiber optic sensor (DFOS) enable distributed strain sensing (DSS) with high spatial resolution over extended length and provide unprecedented opportunities for structural health monitoring (SHM) of infrastructure. However, automated evaluation of the data – the high resolution results in a large data volume – still poses a challenge. Especially measurement/related disturbances – such as strain reading anomaly (SRAs), dropouts, and noise – impede automated evaluation as they may cause serious misinterpretation. For civil engineers, the primary interest in DFOS data lies in extracting reliable information on structural integrity, which necessitates effective preprocessing to eliminate these disturbances. This study presents common types of DFOS data disturbances, their underlying causes, and selected algorithms for automated preprocessing. The capabilities of the proposed algorithms are demonstrated through benchmark tests, providing practical recommendations for improving the accuracy and reliability of DFOS data evaluation in SHM applications.