This　paper　proposes　an　active　disturbance　rejection　predictive　control　(ADRPC)　strategy　to　improve　the　robustness　of　the　six-phase　permanent　magnet　synchronous　machine　(PMSM)　regulation　system　under　complex　working　conditions.　Compared　to　the　traditional　linear　active　disturbance　rejection　control　(LADRC),　ADRPC　adopts　the　generalized　predictive　control　(GPC)　for　the　linear　state　error　feedback　control　(LSEF)　law　to　improve　the　control　performance.　Due　to　the　continuous　moving　optimization　based　on　the　minimum　variance　control,　this　law　improves　the　speed　response　compared　to　the　conventional　LSEF　while　maintaining　stability.　The　system　model　and　the　controller　mathematical　derivation　equation　are　established　in　this　paper　to　build　the　simulation　model.　It　is　verified　that　ADRPC　has　better　robustness　than　LADRC　according　to　the　comparison　of　simulation　results.