[Objective] To address the issue of reduced servo accuracy in permanent magnet synchronous linear motors (PMSLM) caused by external load disturbances and parameter perturbations due to temperature rise during operation, this paper employs a sliding mode controller (SMC) based on the super-twisting algorithm (STA) for the position loop. A dual online iterative compensation (DOIC) control strategy, namely STA-DOIC, is proposed. [Methods] Firstly, a predictive framework was constructed based on the discretized PMSLM mathematical model, and a load disturbance observer (LDO) and a parameter disturbance observer (PDO) were introduced to accurately estimate external load variations and internal parameter perturbations. Second, the compensation timing was dynamically planned within the prediction horizon, and the optimal compensation amount was solved online based on the iterative learning algorithm. Finally, the compensation amount was injected into the control system in real time, effectively eliminating the residual error of the STA and significantly improving tracking accuracy. [Results] Based on simulation and experimental data, the STA-DOIC control strategy demonstrated superior performance under various operating conditions, including no-load, load, and parameter mismatch, with average tracking error reductions exceeding 95% (simulation) and 99% (experiment). The introduction of LDO and PDO further enhanced the system’s disturbance rejection capability, reducing errors under both external load and parameter perturbations. The parameter η=3,600 was determined as the optimal value, ensuring high precision while maintaining system stability. [Conclusion] The STA-DOIC control strategy proposed in this paper, combined with PDO and LDO, effectively enhances the system's robustness against disturbances and parameter uncertainties, providing reliable assurance for high-precision servo control.