When tracking noisy signals with traditional adaptive tracking differentiator,disturbance may be introduced into rate factor and filtering factor of the differentiator and leads to disorder and tracking to the noise.To solve this problem,the cause of noise was analyzed,performances of adaptive tracking differentiators were studied,and an Improved Adaptive Tracking Differentiator (IATD) was proposed.By using the adaptation function taking the Moving Window Standard Deviation (MWSD) estimation of track spacing as input,the IATD could adjust the rate factor and filtering factor optimally.Simulation indicates that IATD can adjust factors more accurately and rapidly with noisy input,and has better noise restraining ability in wide range of signal noise ratio.