Self-mixing interferometry (SMI) signals are observed from a sensing system consisting of a laser diode with external optical feedbacks. SMI signals carry the information associated to both of the displacement and parameters of the SL. To retrieve the information precisely, pre-processing of SMI signals is the first key step. For achieving real-time, and high quality sensing, this paper proposes a Field-programmable gate arrays (FPGA) based filtering and normalizing processing for SMI signals. According to the noise features contained in SMI signal, a median filter and a wavelet transform based filter are combined for our design. Hardware co-simulation verified that the performance for this FPGA used filter design.