A new result for the analysis of positive Markov jump neural networks(PMJNNs) with time delay is described in this paper. By rewriting the PMJNNs with time delay in both continues-time and discrete-time domains into equivalent positive neural networks(PNNs) and analyzing their stability issues, two delay-dependent sufficient conditions are presented to ensure that the continuous-time and the discrete-time PMJNNs with time delay are exponentially mean stable(EMS) through using the inequality technique. All conditions obtained in the paper are in terms of standard linear programming, which reduces the conservatism. Finally, two numerical examples are provided to verify the validity of our results.