Increased cortical porosity is a major cause of the decreased strength of osteoporotic bone, which can be evaluated by MRI based bone water components analysis. The chemical shift caused by fat in the bone may lead to incorrect estimation of water components with a bi-component exponential model. Thus, we propose a tri-component fitting method for accurate bound and pore water quantification incorporating a multi-peak spectral modeling of fat. Nine human cortical bone samples were studied. Our results suggest improved curve fitting and additional information when a tri-component analyses is used.