The mismatch between training and testing environments makes the necessity of speech recognizers to be adaptive both in acoustic modeling and decision rule. Accordingly, the speech hidden Markov models (HMM's) should be able to incrementally capture the evolving statistics of environments. Also, the speech recognizer should incorporate the inevitable parameter uncertainty for robust decision. This paper presents a transformation-based Bayesian predictive classification where the uncertainties of transformation parameters of HMM mean vector and precision matrix are adequately represented by a conjugate prior density. Due to the benefit of conjugate density, we generate the reproducible prior/posterior pair such that the hyperparameters of prior density could be evolved successively to new environments using online test data. The evolved hyperparameters could suitably describe the parameter uncertainty for TBPC decision. Therefore, a novel framework of TBPC geared with online prior evolution is developed for robust speech recognition. This framework is examined to be effective and efficient on the recognition task of connected Chinese digits in hands-free car environments.