Oil and gas pipelines crossing complex geological regions are often subjected to additional loads induced by soil movement during service, resulting in pipeline deflection and deformation. As critical components of oil and gas pipelines, abnormal bending strain data based on IMU (inertial measurement unit) can be caused by bends, which is considered to affect the accurate assessment of pipeline bending deformation states. The positions and geometric characteristics of bends are aimed to be accurately identified based on pipeline centerline data, providing technical support for pipeline bending strain analysis and safety assessment. Based on the pipeline centerline coordinates obtained from IMU, the cubic spline interpolation method was applied to segmentally fit the eastward, northward, and upward trends of the pipeline. The curvature was calculated to determine the locations and geometric characteristics of the bends. The proposed method was verified using the centerline data and MFL (magnetic flux leakage) results from a 4.4 km long-distance pipeline. The error in bend start and end positioning is controlled within 0.025 m by the proposed bend identification method. For detection sensitivity, small bends that cannot be detected by MFL are identified by the proposed method, and the total number of bends identified is 8.0% higher than that detected by MFL. Regarding detection reliability, a recognition accuracy of 96.4% is achieved by this method among the 137 bends detected by MFL. The positions and angles of bends can be quickly and accurately located by the proposed bend identification method based on pipeline centerline coordinate data. The method is applicable to the identification of both hot and cold bends. The analysis of abnormal pipe bending strain measurements based on inertial measurement units is supported by this method, and technical support is provided for constructing of three-dimensional digital models of aging pipelines.