BACKGROUND: The primary stability of a cementless implant is crucial to ensure long term stability through osseointegration. In the present study we have examined how subject specific finite element models can be used to evaluate the stability of a cementless femoral stem. METHODS: Micromotion on the bone-implant interface of a cementless stem was measured experimentally in six human cadaver femurs. Subject specific finite element models were built from computed tomography of the same femurs, and used to simulate the same load scenario used experimentally. FINDINGS: Both experimental measurements and numerical analyses showed a tendency of increased rotational stability for bigger implants. Good correlation was found between measurements and calculated values of axial rotation (R(2)=0.74, P<0.001). The finite element models produced interface micromotion of the same magnitude as measured experimentally, with micromotion generally below 40 microm. Bigger femoral stems were found to decrease the micromotion in the experimental measurements. This tendency could not be recognised in the interface micromotion from the finite element models. INTERPRETATION: The finite element models showed limited success in predicting interfacial micromotion, but reproduced a similar pattern of rotational stability for the implants as seen experimentally. Since rotation in retroversion is often the main concern when studying implant stability, subject specific finite element models could be employed for pre-clinical evaluation of implants.