In this paper we present a semantic video indexing algorithm based on finite-state machines and low-level motion indices extracted from the MPEG compressed bit-stream. The problem of semantic video indexing is actually of great interest due to the wide diffusion of large video databases. In literature we can find many video indexing algorithms, based on various types of low-level features, but the problem of semantic indexing is less studied and surely it is a great challenging one. The proposed algorithm is an example of solution to the problem of finding a semantic relevant event (e.g., scoring of a goal in a soccer game) in case of specific categories of audio-visual programmes. The simulation results show that the proposed algorithm can effectively detect the presence of goals and other relevant events in sport programs.