In this paper, the link strength of each network switch in Software Defined Networking (SDN) environment system is analyzed into either good or bad using the proposed deep learning approach. This proposed method is designed with two modules as feature computation and Convolutional Neural Network (CNN) architecture. The feature computation module computes the intrinsic feature values of each network switch in SDN, and these intrinsic features are trained and classified into either good or bad using the proposed CNN architecture. The performance of the proposed SDN system is analyzed using the parameters precision, recall, accuracy, average detection rate, Packet Delivery Ratio (PDR) and latency.