AIM: To determine a valid and reliable neck-shaft angle (NSA) measurement method while rotating the pelvises in increments of 5° in order to simulate patient malpositioning. MATERIALS AND METHODS: CT images of 17 patients were used to produce digitally reconstructed radiographs in frontal and lateral views and three-dimensional (3D)-reconstructions of the femurs, considered to be the reference standard. Malpositioning was simulated by axially rotating the frontal radiographs from 0° to 20°. Three operators measured in two-dimensions the NSA using four different methods, three times each, at each axial rotation (AR) position. Method 1 (femoral neck axis drawn by joining the centre of the femoral head (CFH) to the median of the femoral neck base; femoral diaphysis axis drawn by joining the median of two lines passing through the medial and lateral edges of the femoral axis below the lesser trochanter) and method 2 (femoral axis taken as the median of a triangle passing through base of femoral neck and medial and lateral head-neck junction; femoral diaphysis as previous) were described for the first time; method 3 was based on a previous study; method 4 was a free-hand technique. Reliability, validity, and global uncertainty were assessed. RESULTS: Method 1 showed the best reliability and validity. The global uncertainty also showed minimal values for method 1, ranging from 7.4° to 14.3° across AR positions. CONCLUSION: Method 1, based on locating the CFH, was the most reliable and valid method and should be considered as a standardised two-dimensional NSA measurement method for clinical application.