Multilayer Perceptron (MLP) with BackPropagation learning algorithm is effective in solving a wide range of bio-medical classification problems. In this paper, the MLP network trained with backpropagation algorithm is implemented on a cluster of workstations using MPI for increased computational speed and effective resource utilization. The proposed work combines data session and training set parallelism to improve the convergence rate and efficiency. This work uses the standard enhancement techniques like momentum factor, adaptive learning rate and adaptive learning rate with momentum factor combined with the standard gradient descent algorithm. The performance of parallel backpropagation algorithm is evaluated for the liver disorder diagnosis and heart disease diagnosis applications. Experimental performance shows that the proposed parallel algorithm has better speedup than the sequential algorithm.