Cognitive radio technology inherently possesses self adaptivity. In order to design self adaptive mobile cognitive radio networks, routing is one of the key challenging issues to be addressed. In this paper, a novel self adaptive routing (SAR) algorithm for multi-hop cognitive radio ad hoc networks is proposed. The proposed routing algorithm incorporates with routing metrics and autonomous distributed adaptive transmission range control mechanism to provide self adaptivity. SAR aims to choose optimal routes at the outset of routing and aims to retain optimal route by the use of route adaptation and route preservation. SAR is compared with previously suggested algorithms to indicate performance differences. Extensive experimental evaluations are performed in the ns2 simulator. It is shown that the SAR provides better adaptability to the environment than the previously suggested algorithms and maximizes throughput, minimizes end-to-end delay in a number of realistic scenarios and significantly improves routing performance. (c) 2013 Elsevier Ltd. All rights reserved.