Abstract: In the high-speed sampling, the parameters of the discrete model based on delta operator tend to the original continuous-time model, and the system has a better digital characteristic when delta operator achieves, which can improve the problem of filtering divergence of discrete Kalman filters based on traditional shift operator. This article uses orthogonal projection approach to derive the Kalman filter equation of stochastic linear discrete systems based on delta operator. The recursive Kalman filter algorithm based on delta operator is given. The simulation results show that in the high-speed sampling, the convergence performance of the derived Kalman filter based on backward differentiation formula of delta operator is better than the traditional Kalman filters.