In application of wireless sensor networks（ WSNs）

effective estimation for link quality is a basic issue in guarantying data reliable transmission and upper layer network protocol performance. Aiming at problem that in existing wireless link quality estimation study

link quality hierarchy division has no uniform standard and models are lack of environmental adaptability

a link quality evaluation mechanism based on Bayesian networks is proposed. As link quality has multiple attributes

close degree analysis method is used for link quality grade division

and use Bayesian networks for link quality uncertainty reasoning and estimation modeling. Through training and testing in multi-application scenarios

practical testing results demonstrate that the link quality estimation model based on Bayesian network has high accuracy and generalization ability in different application environments

reflect good adaptability.