There are many advantages of direct nose-to-brain drug delivery in the treatment of neurological disorders. However, its application is limited by the extremely low delivery efficiency (< 1%) to the olfactory mucosa that directly connects the brain. It is crucial to develop novel techniques to deliver neurological medications more effectively to the olfactory region. The objective of this study is to develop a numerical platform to simulate and improve intranasal olfactory drug delivery. A coupled image-CFD method was presented that synthetized the image-based model development, quality meshing, fluid simulation, and magnetic particle tracking. With this method, performances of three intranasal delivery protocols were numerically assessed and compared. Influences of breathing maneuvers, magnet layout, magnetic field strength, drug release position, and particle size on the olfactory dosage were also numerically studied. From the simulations, we found that clinically significant olfactory dosage (up to 45%) were feasible using the combination of magnet layout and selective drug release. A 64 -fold higher delivery of dosage was predicted in the case with magnetophoretic guidance compared to the case without it. However, precise guidance of nasally inhaled aerosols to the olfactory region remains challenging due to the unstable nature of magnetophoresis, as well as the high sensitivity of olfactory dosage to patient-, device-, and particle-related factors.