During the operation of a checkweigher, its weighing signal is affected by vibrations rising from the mechanical transmission systems, impacts from the measured object and other random disturbances. As a result, the weighing signal is severely contaminated by noise, making it difficult to meet the requirements of national standards. To address this issue, a novel dynamic weighing method based on improved stationary wavelet denoising with a shrinkage soft threshold and extended Kalman system identification is proposed. First, leveraging prior knowledge of the weighing signal and the ideal signal, a seven-layer stationary wavelet transform is applied to the weighing signal for multi-scale decomposition. Next, the high-frequency noise-dominated detail coefficients d1,k~d4,k are set to zero, while a soft-threshold function with a shrinkage factor is applied to process the detail coefficients d5,k~d7,k that contain both useful signal and noise components. Then, the inverse stationary wavelet transform is performed using the processed detail coefficients and the original approximation coefficients to reconstruct the weighing signal, effectively suppressing various interference noises. Following this, the extended Kalman algorithm is employed for system identification to determine the model parameters of the checkweigher system, which are subsequently utilized to calculate the mass of the samples. To validate the effectiveness of the proposed algorithm, experiments were conducted using five samples of different masses at speeds of 30, 45, 60, 75 and 90 m/min, with multiple loading tests performed at each speed, and the results were analyzed and compared. The results demonstrate that the proposed algorithm achieves superior weighing accuracy compared to the time-variant low-pass filter (TVLPF) algorithm, identification-based approach with signal-adaptive prefiltering (AID) algorithm, and signal-adaptive prefiltering with extended Kalman system identification (AEKSI) algorithm. Furthermore, it meets the accuracy requirements for class XIII checkweighers as defined by the national standard “GB/T 27739—2011 Automatic Checkweigher”.