The large size of nanoparticles prevents rapid extravasation from blood vessels and diffusion into tumors. Multimodal imaging uses the physical properties of one modality to validate the results of another. We aim to demonstrate the use of a targeted thin layer-protected ultra-small gold nanoparticles (Au-NPs) to detect cancer in vivo using multimodal imaging with photoacoustic and computed tomography (CT). The thin layer was produced using a mixed thiol-containing short ligands, including MUA, CVVVT-ol, and HS-(CH2)11-PEG4-OH. The gold nanoparticle was labeled with a heterobivalent (HB) peptide ligand that targets overexpression of epidermal growth factor receptors (EGFR) and ErbB2, hereafter HB-Au-NPs. A human xenograft model of esophageal cancer was used for imaging. HB-Au-NPs show spherical morphology, a core diameter of 4.47 ± 0.8 nm on transmission electron microscopy, and a hydrodynamic diameter of 6.41 ± 0.73 nm on dynamic light scattering. Uptake of HB-Au-NPs was observed only in cancer cells that overexpressed EGFR and ErbB2 using photoacoustic microscopy. Photoacoustic images of tumors in vivo showed peak HB-Au-NPs uptake at 8 h post-injection with systemic clearance by ~48 h. Whole-body images using CT validated specific tumor uptake of HB-Au-NPs in vivo. HB-Au-NPs showed good stability and biocompatibility with fast clearance and contrast-enhancing capability for both photoacoustic and CT imaging. A targeted thin layer-protected gold nanoprobe represents a new platform for molecular imaging and shows promise for early detection and staging of cancer.