The paper studies the effect of high-velocity oxy-fuel thermal spraying parameters, in particular spray distance and oxygen flow rate, on coating porosity and magnetic properties of FeSi and FeSiB deposits using the artificial neural network methodology. The magnetic properties correlated to coating porosity were obtained using an optimized network structure. The predicted results permitted to point out the role of porosity for varying the coercivity and saturation magnetization and the stability of magnetic properties with respect to the considered spray parameters. © 2005 Elsevier B.V. All rights reserved.