Remote control and targeted activation are the elevated goals for shape-memory and self-healing polymers when responses to stimuli. Among so many stimuli, converting light into microstructure change or mechanical motions is now of particular interest. Herein, we report the ingenious design, synthesis and operation to advanced materials that capable of fast near-infrared (NIR) light-actuated targeted shape memory and remote accurately self-healing. Starting from biomass resources, a well-defined polymer nanocomposite, CNTs-graft-poly(tetrahydrofurfury methacryla-co-lauryl acrylate-co-1-vinylimidazole) copolymer ((CNTs-g-P(TMA-co-LA-co-VI)), was fabricated by addition-fragmentation chain transfer (RAFT) polymerization. After subsequent metal-ligand crosslinking with Zn2+ ion, CNTs-g-P(TMA-co-LA-co-VI)/Zn2+ with CNTs content of 1.1 wt % have a maximum stress of 1.68 MPa and an elongation at break of 450%. Most importantly, the photothermal conversion of CNTs can effectively trigger the association and dissociation process of the dynamic metallosupramolecular crosslinked bond between VI and Zn2+, leading to excellent instantaneous multiple shape memory and accurate self-healing performance under NIR light or heat. This approach can be generalized toward de novo design of self-healing and shape memory nanocomposites with tunable mechanical properties. Display omitted •Starting from biomass resources, a well-defined elastomer composite was fabricated by RAFT polymerization.•The elastomer showed excellent multiple shape memory and accurate self-healing performance under NIR light or heat.•This multifunctional elastomer was applicable to the fabrication of smart devices and soft robotics.