Abstract Focused on the common problem of easy specimen buckling in the axial strain–controlled fatigue testing of thin automobile steel sheets, a strain fatigue test method for thicknesses less than 2.5 mm is proposed. Because of the unrestrained grip end of specimens between the anti-buckling device and the testing machine fixture, specimens were prone to buckling in this area. In addition, if the designed specimen sizes were unreasonable, buckling may have occurred in the width and thickness directions of the gauge length of the specimen. Both situations can lead to invalid tests. This paper adopted a testing machine fixture and an anti-buckling plate as a mortise-and-tenon structure for the anti-buckling method; the upper and lower fixtures of the testing machine were cut in straight grooves, and the left and right anti-buckling plates were designed to be connected with a mortise and tenon. This is equivalent to narrowing the width of the unrestrained part of the specimen’s grip end to avoid buckling failure due to plane strain. The factors affecting the stiffness of fatigue specimens are also discussed herein. In order to avoid common instability in the thickness and width directions of the gauge length and unconstrained part of the specimen’s grip end at a strain ratio of Re = −1 during testing, six recommended specimen sizes are given on the premise of ensuring maximum specimen stiffness. When using the anti-buckling device, the recommended specimen sizes and test points of this study (i.e., the lower limits for yield strength and thickness of strain control testing) are 110 MPa and 0.5 mm, respectively, with which continuous and smooth stress–strain curves that result in accurate and reliable experimental data can be obtained.