This　paper　concentrates　on　the　global　predefined-time　synchronization　of　delayed　memristive　neural　networks　with　external　unknown　disturbance　via　an　observer-based　active　control.　First,　a　global　predefined-time　stability　theorem　based　on　a　non-negative　piecewise　Lyapunov　function　is　proposed,　which　can　obtain　more　accurate　upper　bound　of　the　settling　time　estimation.　Subsequently,　considering　the　delayed　memristive　neural　networks　with　disturbance,　a　disturbance-observer　is　designed　to　approximate　the　external　unknown　disturbance　in　the　response　system　with　a　Hurwitz　theorem　and　then　to　eliminate　the　influence　of　the　unknown　disturbance.　With　the　help　of　global　predefined-time　stability　theorem,　the　predefined-time　synchronization　is　achieved　between　two　delayed　memristive　neural　networks　via　an　active　control　Lyapunov　function　design.　Finally,　two　numerical　simulations　are　performed,　and　the　results　are　given　to　show　the　correctness　and　feasibility　of　the　predefined-time　stability　theorem.