The geometrical shape of the femoral component of hip implants plays a key role in the long-term survivorship of hip implants. The aim of this study was to propose a multi-objective shape optimization procedure using the MOPSO algorithm with three shape-dependent failure mechanisms of hip implants as objective functions including the stress shielding, initial relative micro-motion, and bone-implant interface stress. The Taperloc® Complete femoral stem was selected and its reference geometry was defined with sixty-seven variables. Ten new stem shapes were produced as the swarm members by randomly changing the values of the variables. The values of the three objectives for each stem shape were calculated by the finite element analysis and the position of each swarm member was updated iteratively using the MOPSO algorithm. The geometry that caused a 37% and 33% decrease in the interface stress and stress shielding, respectively, and a 32% increase in the initial micro-motion compared to the Taperloc® Complete stem was selected as the optimized shape. It was shown that thinning the femoral stems without changing their length reduced the induced stress shielding and initial micro-motion and increased the interface stress, whereas shortening the femoral stems reduced the stress shielding and interface stress and increased the initial micro-motion. The proposed procedure may be conveniently used for the shape optimization of commercial femoral stems, which may significantly impact the performance and lifetime of hip implants.