The aim of this study was to evaluate the effects of using a portable, accelerometer-based surgical navigation system (KneeAlign2) in total knee arthroplasty (TKA) on the alignment of the femoral component, and blood loss. A total of 241 consecutive patients with primary osteoarthritis of the knee were enrolled in this prospective, randomised controlled study. There were 207 women and 34 men. The mean age of the patients was 74.0 years (57 to 89). The KneeAlign2 system was used for distal femoral resection in 121 patients (KA2 group) and a conventional intramedullary femoral guide was used in 120 patients (IM group). One patient (0.8%) in the KA2 group and 19 in the IM group had an alignment which was > 3° away from the neutral mechanical axis (p < 0.01). The mean deviation from neutral alignment was 1.01° (standard deviation (sd) 1.0°) in the KA2 group and 1.93° (sd 1.7°) in the IM group (p < 0.01). Blood loss was significantly less in the KA2 group compared with the IM group (784 ml (sd 357) versus 1071 ml (sd 310), p < 0.001). The KneeAlign2 system provides a technically straightforward method for identifying the femoral head and performing an accurate distal femoral resection at TKA with significantly less blood loss compared with a conventional intramedullary guide. Cite this article: Bone Joint J 2017;99-B:1047-52.