Abstract: Convolutional coding is a forward error correction technique. Viterbi algorithm, which bases upon Maximum-Likelihood estimation, is an outstanding decoding algorithms used with convolutional code. Viterbi algorithm is suited for hardware implementation. In this thesis, we concentrate on how to implement a Viterbi decoder with DSP. To test the decoder, we take a BMP file as the data source, and simulate an AWGN channel in software. This can be widely used in digital communications applications.