This paper presents a set of ‘strategies’ that enabled the development of a real-time continuous speech recognition system for Czech language. The optimization strategies include efficient computation of HMM probability densities, pruning schemes applied to HMM states, words and word hypotheses, a bigram compression technique as well as parallel implementation of the real recognition system. In a series of off-line speaker-independent tests done with 1600 Czech sentences based on 7033-word lexicon we got 65 % recognition rate. Several on-line tests proved that similar rates can be achieved under real conditions and with response time that is shorter than 1 second.