Railway bridges are the expensive and key element of railway infrastructure and have consequences when they fail or when their capacity is impaired. Ensuring the safe operation of these railway bridges is of importance. The current practice of visual inspection and non-destructive evaluation techniques are more often time consuming and difficult to perform in places that are inaccessible. Condition monitoring aims at measuring some kinematic responses such as displacements, strains, and/or accelerations at different locations of the bridge under normal operating conditions and uses this information to determine different parameters such as boundary conditions, system connectivity, material, and geometrical properties, which help in the estimation of stiffness of the structure. Strain and acceleration responses are commonly used to monitor the condition of a bridge. Out of these two, strain responses are useful because of their ability to give information about the stresses in the structure during its operation, fatigue, or yielding of the material. Reduction in the axial stiffness of the members is one of the anomalies in truss bridges. An anomaly identification algorithm is proposed that uses anomaly-sensitive features derived from strain responses for a statically determinate steel railway truss bridge. To detect the changes in stiffness, mechanics-based and data-driven approaches are studied. Locomotive axle load, localized cross-sectional area, and root mean square error between field strain responses and analytically estimated strains are the anomaly features considered for the mechanics-based framework. Field-measured strain responses and anomaly feature derived using principal component analysis of the field-measured strain responses are the anomaly features considered for the data-driven based framework. The non-stationary nature of these anomaly features is used to detect abnormalities. The anomaly identification framework is validated with the strain responses from a real-life steel truss railway bridge.