Since　the　inverse　model　is　a　key　technique　and　difficult　to　design,　a　novel　method　is　proposed　to　compute　the　inverse　model　for　internal　model　control　by　Newton　method.　Specifically,　support　vector　machine　estimates　the　object　as　the　internal　model.　Then,　the　Newton　method　solves　the　root　of　the　internal　model,　as　the　inverse　model.　Moreover,　the　stability　is　analyzed　for　the　proposed　system.　This　strategy　is　used　to　control　permanent　magnet　linear　synchronous　motor　in　the　simulations.　The　results　show　that　since　the　inverse　model　by　the　Newton　method　well　matches　the　internal　model,　the　proposed　system　for　permanent　magnet　linear　synchronous　motor　possesses　the　performance　of　a　higher　tracking　position　accuracy,　a　stronger　robustness,　compared　with　the　conventional　internal　model　control.　©　2015　ISSN　2185-2766.