PurposeWith　the　evolving　disease　spectrum,　chronic　diseases　have　emerged　as　a　primary　burden　and　a　leading　cause　of　mortality.　Due　to　the　aging　population　and　the　nature　of　chronic　illnesses,　patients　often　suffer　from　multimorbidity.　Predicting　the　likelihood　of　these　patients　developing　specific　diseases　in　the　future　based　on　their　current　health　status　and　age　factors　is　a　crucial　task　in　multimorbidity　research.MethodsWe　propose　an　algorithm,　CE-GCN,　which　integrates　age　sequence　and　embeds　Graph　Convolutional　Network　(GCN)　into　Gated　Recurrent　Unit　(GRU),　utilizing　the　topological　feature　of　network　common　neighbors　to　predict　links　in　dynamic　complex　networks.　First,　we　constructed　a　disease　evolution　network　spanning　from　ages　45　to　90　years　old　using　disease　information　from　3333　patients.　Then,　we　introduced　an　innovative　approach　for　link　prediction　aimed　at　uncovering　relationships　between　various　diseases.　This　method　takes　into　account　patients＇　age　to　construct　the　evolutionary　structure　of　the　disease　network,　thereby　predicting　the　connections　between　chronic　diseases.ResultsResults　from　experiments　conducted　on　real　networks　indicate　that　our　model　surpasses　others　regarding　both　MRR　and　MAP.　The　proposed　method　accurately　reveals　associations　between　diseases　and　effectively　captures　future　disease　risks.ConclusionOur　model　can　serve　as　an　objective　and　convenient　computer-aided　tool　to　identify　hidden　relationships　between　diseases　in　order　to　assist　healthcare　professionals　in　taking　early　disease　interventions,　which　can　substantially　lower　the　costs　associated　with　treating　multimorbidity　and　enhance　the　quality　of　life　for　patients　suffering　from　chronic　conditions.