Conventional I–f startup control algorithms often lead to significant speed fluctuations, extended convergence times, and sluggish dynamic responses when the motor is working at the acceleration process. To address these challenges, the article constructs a discrete kinetic model of the motor, extracts and summarizes the change law of the motor torque angle fluctuation amplitude and frequency with the motor parameters and the reference current vector during I–f starting process based on the Runge–Kutta method, reveals the fundamental mechanism of the pole slipping of the I–f control. On this basis, the article proposes a closed‐loop I–f starting control strategy for high‐speed permanent magnet motors based on current vector adaptive regulation, The strategy notably improves speed tracking performance and current dynamic response to I–f startup control. Experimental results validate the efficacy of the proposed algorithm.