Impulse noise is one of the important facts impairing the performance of the power line communications systems (PLC). The partitioned Markov chain was proposed to evaluate the impairments caused by the impulse noise and to model the power communications channels. The Baulm- Welch algorithm was employed to calculate the basic transition matrix to avoid the less robust computation with the nonlinear fitting method. In order to easy and quick computation, the partitioned Markov chain was proposed to evaluate the Baulm-Welch algorithm has been adapted to the partitioned Markov chain. To verify the proposed method, four modulation schemes have been implemented in both home and office environments. Based on the measurements, the corresponding basic transition matrixes were calculated. The detailed analysis of the basic transition matrixes shows the good agreement with other research results, which mean the correctness of the proposed partitioned Markov chain to model PLC channels