<p>This study proposes an optimal model for reducing the cogging force of a permanent magnet linear synchronous motor. Among conventional methods for cogging reduction, both linear and step skew methods were examined to evaluate their effects on thrust and cogging force. Previous studies simply compared the electromagnetic performances when linear skew and 2-step skew were applied, this work investigates multi-step configurations and provides an analytical interpretation of their non-monotonic behavior using the Dirichlet kernel concept. The analysis clarifies how segment division influence cogging reduction and thrust force degradation. Also, this study explained the trade-off between electromagnetic smoothness and manufacturing complexity. Based on this comparison, one of the skew methods is selected and applied for optimization to design the optimal model. A genetic algorithm was employed, with both thrust and cogging force set as objective functions. To prevent excessive magnet usage due to skew applications, the magnet volume was constrained to within a 5% increase. The optimal model achieved a 0.5% improvement in thrust force relative to the requirement, a 94.56% reduction in cogging force, and satisfied the magnet volume constraint.</p>