Mechanical stiffness is a fundamental parameter in the rational design of composites for bone tissue engineering in that it affects both the mechanical stability and the osteo-regeneration process at the fracture site. A mathematical model is presented for predicting the effective Young's modulus (E) and shear modulus (G) of a multi-phase biocomposite as a function of the geometry, material properties and volume concentration of each individual phase. It is demonstrated that the shape of the reinforcing particles may dramatically affect the mechanical stiffness: E and G can be maximized by employing particles with large geometrical anisotropy, such as thin platelet-like or long fibrillar-like particles. For a porous poly(propylene fumarate) (60% porosity) scaffold reinforced with silicon particles (10% volume concentration) the Young's (shear) modulus could be increased by more than 10 times by just using thin platelet-like as opposed to classical spherical particles, achieving an effective modulus E approximately 8 GPa (G approximately 3.5 GPa). The mathematical model proposed provides results in good agreement with several experimental test cases and could help in identifying the proper formulation of bone scaffolds, reducing the development time and guiding the experimental testing.