An acoustic emission signal processing system based upon a $\mu$C has been designed, constructed, and tested. The system consists of a piezoelectric transducer, low-noise wide-band preamplifier, peak detector, analog-to-digital converter, FIFO buffer memory, $\mu$C-based amplitude sorter, display interface, and oscilloscope. The system detects the local-peaks of a series of AE pulse signals and classifies them into eighty different levels, from 1 dB to 80 dB, and counts the events $N_i$ at each level $V_i$. Measured results are displayed on an oscilloscope in two formats: (1) log $V_i$ versus log Ni,(2) log $V_i^2$ versus log $N_iV_i^2$; both of them give important informations in the study of AE phenomenon. The designed system, easy to operate, economic, and flexible, will be further developed and used in the laboratory for the study of AE phenomenon in the near future.