Background Accurate repeat assessment of the diameter of an abdominal aortic aneurysm (AAA) is important. This study investigated the reproducibility of different methods of measuring AAA diameter from ultrasound images. Methods Fifty AAA patients were assessed by ultrasound. Maximum AAA diameter was measured independently by three trained observers on two separate occasions using a standardised protocol. Five diameters were measured from each scan, three in the anterior–posterior (AP) and two in the transverse (TV) plane, including inner-to-inner (ITI), outer-to-outer (OTO) and leading edge-to-leading edge (LETLE). Intra- and inter-observer reproducibility were reported as reproducibility coefficients. Statistical comparison of methods was performed using linear mixed effects models. Results Intra-observer reproducibility coefficients (AP LETLE 2.2 mm; AP ITI 2.4 mm; AP OTO 2.6 mm) were smaller than inter-observer reproducibility coefficients (AP LETLE 4.6 mm: AP ITI 4.5; and AP OTO 4.8 mm). There was no statistically significant difference in intra-observer reproducibility of three types of measurements performed in the AP plane. Measurements obtained in the TV plane had statistically significant worse intra-observer reproducibility than those performed in the AP plane. Conclusions This study suggests that the comparison of maximum AAA diameter between repeat images is most reproducibly performed by a single trained observer measuring diameters in the AP plane.