Anticounterfeiting　technologies　meet　challenges　in　the　Internet　of　Things　era　due　to　the　rapidly　growing　volume　of　objects,　their　frequent　connection　with　humans,　and　the　accelerated　advance　of　counterfeiting/cracking　techniques.　Here,　we,　inspired　by　biological　fingerprints,　present　a　simple　anticounterfeiting　system　based　on　perovskite　quantum　dot　(PQD)　fingerprint　physical　unclonable　function　(FPUF)　by　cooperatively　utilizing　the　spontaneous-phase　separation　of　polymers　and　selective　in　situ　synthesis　PQDs　as　an　entropy　source.　The　FPUFs　offer　red,　green,　and　blue　full-color　fingerprint　identifiers　and　random　three-dimensional　(3D)　morphology,　which　extends　binary　to　multivalued　encoding　by　tuning　the　perovskite　and　polymer　components,　enabling　a　high　encoding　capacity　(about　108570000,　far　surpassing　that　of　biometric　fingerprints).　The　strategy　is　compatible　with　mainstream　production　techniques　that　are　widely　used　in　traditional　low-cost　printed　anticounterfeiting　labels　including　spray　printing,　stamping,　writing,　and　laser　printing,　avoiding　complicated　fabrication.　Macrographical　patterns　and　micro/nanofingerprint　patterns　with　multiscale-tailorable　inter-ridge　sizes　can　be　fused　into　a　single　FPUF　label,　satisfying　different　levels　of　anticounterfeiting　requirements.　Furthermore,　a　smart　fused　scheme　of　enhanced　deep　learning　and　fingerprint　characteristic　comparison　is　leveraged,　by　which　high-efficiency,　high-accuracy　authentication　of　our　FPUFs　is　achieved　even　for　the　increasingly　huge　FPUF　databases　and　imperfectly　captured　images　from　users.