This paper proposes a parameter generation algorithm using lo-cal variance (LV) constraint of spectral parameter trajectory for HMM-based speech synthesis. In the parameter generation pro-cess, we take account of both the HMM likelihood of speech feature vectors and a likelihood for LVs. To model LV precisely, we use dynamic features of LV with context-dependent HMMs. The objective experimental results show that the proposed tech-nique can generate a better spectral trajectory in terms of the spectral and LV distortions than a conventional technique with global variance (GV) constraint. The subjective experimental results also show that the proposed technique significantly im-prove the reproducibility of the synthetic speech than the con-ventional one.