Aiming to address the issues of low weighing accuracy and suboptimal efficiency caused by multiple disturbances during the weighing process of tea packaging machines, focusing on weighing signal denoising and controling algorithm optimization, a comprehensive method integrating an optimizing Kalman filter algorithm with a three-stage fuzzy PD and adaptive iterative learning control strategy was proposed. Regarding signal denoising, addressing the insufficient noise reduction effectiveness of traditional Kalman filtering across weighing stages, a phased optimization strategy was proposed: during the dynamic feeding stage, exponential preprocessing was fused with Kalman filtering to suppress highfrequency noise;during static weighing, process disturbance noise covariance was reduced to enhance filtering stability, and final weight values were obtained through weighted averaging based on the convergence degree of the state covariance matrix. In the hopper opening phase, tracking lag was eliminated by adjusting the Kalman gain to its optimum value. During hopper closing, weighted limiting was introduced to mitigate spike disturbances. Regarding the control algorithm, a three-stage fuzzy PD control strategy was designed, dividing the dynamic feeding process into coarse feeding, deceleration feeding, and fine feeding stages. Combining fuzzy theory, PD parameters were tuned online to balance the deceleration stage through dynamic parameter adjustment, facilitating a smooth transition. Furthermore, to address overshoot after the vibrating feeder stops, an adaptive iterative learning algorithm was introduced during the fine feeding stage. By iteratively adjusting the advance stop amount of the vibrating feeder, the actual weighing value converged more rapidly toward the target value. Test results showed the actual weighing deviation for Biluochun and Longjing green teas was controlled within ± 0. 06 g relative to the target mass, while Wuyi rock tea remained within ±0.12g. Moreover, all three tea types completed weighing operations within a short timeframe, with weighing time variations for identical target weights controlled within ±1s. In summary, this method effectively enhanced the precision, efficiency, and stability of the automatic tea weighing system, providing a viable solution for optimizing the performance of automatic tea packaging machines.