In this paper, we describe a training method for continuous mixture density HMMs, named optimal discriminative training (ODT), and its implementation for speech recognition in noise. Conventional maximum likelihood estimation method (MLE) for HMM training has a problem that the recognition performance is not considered in the training procedure. ODT is one of corrective learning methods, applied to continuous mixture density HMMs, and these HMMs are especially useful for speaker-independent speech recognition. Under noisy environments, the recognition categories are tend to be difficult to discriminate in feature space, so by using ODT the improvement of recognition accuracy is more expected. Here, after description the training algorithm of ODT, we discuss the effects of ODT to improve the robustness for adverse environments by the word recognition experiments in noise, where the noise signal takes various levels and spectral characteristics.