Abstract: The deterministic sampling filtering is improved,and a kind of adaptive deterministic sampling filter is proposed. Firstly,for the estimation of noise's statistical parameters,a recursive estimation algorithm for stationary noise and three other ones for non-stationary noise are given;then through combining these estimation algorithms with deterministic sampling filtering,several new adaptive sampling filtering algorithms can be derived.The simulation results are shown that,when statistical parameters of non-stationary noise are unknown,more accurate estimation result can be obtained by adaptive deterministic sampling filtering algorithms.