Event-triggered　control　is　an　effective　control　approach　with　control　update　instants　determined　by　pre-defined　event-triggering　conditions,　and　the　triggering　conditions　are　usually　required　to　be　tested　continuously.　In　this　paper,　we　present　an　adaptive　event-triggered　distributed　model　predictive　control　algorithm　for　nonlinear　continuous-time　multi-agent　systems.　All　the　agents　are　able　to　adaptively　determine　when　to　check　the　triggering　conditions,　which　means　the　triggering　conditions　are　tested　intermittently　with　intervals　determined　adaptively.　We　prove　the　feasibility　of　the　proposed　algorithm,　and　also　prove　that　the　multi-agent　system　under　the　algorithm　will　converge　to　an　invariant　set　in　finite　time　by　utilizing　variable　prediction　horizon　approach　instead　of　the　Lyapunov　function　method,　such　that　the　event-triggering　conditions　are　less　conservative.　Finally,　we　provide　numerical　examples　to　verify　the　effectiveness　and　superiority　of　the　proposed　algorithm,　revealing　that　the　proposed　algorithm　can　efficiently　reduce　the　communication　and　computation　cost　as　well　as　the　sensing　load.　(C)　2019　Elsevier　B.V.　All　rights　reserved.