Image encryption is a field of cryptography where images are encrypted to ensure their safe transmission as well as proper decryption at the receiver’s side. The current research proposes a novel image encryption technique involving chaotic mapping techniques such as Henon maps, Tent maps, and Logistics maps, along with Graph Neural Network (GNN). The technique extracts features from image blocks using GCNConv to exploit graph topology, and later encrypts images by scrambling them block-wise, using chaotic sequences. The results of our encryption technique are compared with preexisting encryption methods using various metrics such as correlation coefficient, information entropy, etc.