The parallel model combination (PMC) technique has been shown to achieve very good performance for speech recognition under noisy conditions. However, there still exist some problems based on the PMC formula. In this paper, we first investigated these problems and some modifications on the transformation process of PMC were proposed. Experimental results show that this modified PMC can provide significant improvements over the original PMC in the recognition accuracies. Error rate reduction on the order of 12.92% was achieved.