It is one of the most important advantages of adaptive filter that it can adjust its parameters automatically according to the change of filtering signal, which makes it possible to adapt to complex signal condition and thereby is used widely in the fields of radar and fuze. The development of DSP (digital signal processor) makes it possible that the filter system based on DSP can fulfill high-efficient and real-time requirements of signal processing. Therefore, in this paper, DSP implementation of adaptive filter is researched and the principle of adaptive filter is analyzed. Taking LMS (Least Mean Square) adaptive filter as an example, the design and implementation of adaptive filter is discussed in detail. Finally, the simulation results on CCS software platform are derived, which validates the correction of this method.