In this proposed method a new routing protocol is SelfAdaptive on demand Geographic routing protocols. One of the distinctive features has provide reliable and energy efficient feature to existing schemes in order to increase the lifetime of the network and reduces control overhead compute up to the proactive schemes which provides reliable routing at the same time with a lesser amount of energy utilization. We are embedding optimization (Marko chain algorithm) on the delivery probability performance in a two-hop relay algorithm with erasure coding. The erasure coding used for the message that is erasure coded into multiple frames at source node. Also we developed finite-state Markov chain model to characterize the complicated message delivery process in the MANETs. Based on the above model, closed-form expressions are further derived for the corresponding message delivery probability under any given message lifetime and message size. The above concept is verified through simulation studies, which can be used to accurately predict the message delivery probability behavior and characterize its relationship with the message size.