Based on the distribution of measurement, a Non-Gaussian model of narrow-band power line noise including background noise and impulsive noise was derived, The background noise, which follows Gaussian distributions, was modeled with a raised cosine roll-off filter, while the impulsive noise was modeled with impulse width, interarrival times and envelope parameters, whose distribution follows partitioned Markov chain. Then, the model was evaluated with some field measurement data and the Markov Chain Monte Carlo (MCMC) analysis. The evaluation results show that the proposed model can approximate the noise in real environment.