The　steady　state　error　caused　by　mismatch　or　changes　of　parameters　in　the　deadbeat　predictive　current　control　(DPCC)　system　of　interior　permanent　magnet　synchronous　motor　is　presented　in　this　paper.　An　improved　model　reference　adaptive　system　(MRAS)　is　proposed.　The　nonlinear　motor　is　viewed　as　reference　model　and　the　adjustable　model　is　discretized　by　the　fourth　order　Runge　Kutta　method.　Inductance　of　d-q　axis.　resistance　and　flux　linkage　are　identified　step　by　step.　The　identified　results　are　brought　into　DPCC　in　real　time　to　eliminate　the　steady　state　error.　Simulation　results　show　that　the　improved　MRAS　can　effectively　solve　the　problem　of　rank　deficiency　and　improve　the　robustness　of　DPCC.　©　2021　IEEE.