For cancer diagnosis, technologies must be capable of molecular recognition, and they must possess a built-in pattern recognition component for efficient imaging and discrimination of targeted cancer cells. Surface enhanced Raman scattering (SERS) tags based on plasmonically active nanoparticles hold promise for accurate and efficient cancer cell recognition, owing to ultra-narrow peak and sensitive optical properties. However, a complex fingerprint spectrum increases data analysis difficulty, making it necessary to develop multicolor SERS tags with a simple fingerprint spectrum. To address this, we herein fabricated SERS-encoded nanoparticles (NPs) with stable and simple fingerprint spectrum through synthesis of isotopic cellular Raman-silent graphene-isolated-Au-nanocrystals (GIANs) and conjugation with phospholipid-polyethylene glycol-linked aptamers to target proteins overexpressed on the cancer cell surface. GIANs, which possess the properties of graphitic nanomaterials, such as super-stable optical properties and high Raman cross-section, showed enhanced SERS signals. The 2D-band Raman shift of GIAN, which located in the cellular Raman-silent region, was easily regulated through fabrication of isotopic GIANs without changing their molecular structure. Such GIAN tags demonstrated multiplexed Raman imaging capability, both in vivo and in vitro, with low background interference. Moreover, cell membrane protein (nucleolin, mucin and epithelial cell adhesion molecule)-specific, aptamer-conjugated isotopic GIANs were fabricated and feasibly applied to built-in coding for rapid imaging and pattern recognition of targeted cancer cells. Such isotopic GIAN-aptamer-encoders show high potential for efficient cancer cell identification and diagnosis.