Arsenic (As) undergoes extensive metabolism in biological systems involving numerous metabolites with varying toxicities. It is important to obtain reliable information on arsenic speciation for understanding toxicity and relevant modes of action. Currently, popular arsenic speciation techniques, such as chromatographic/electrophoretic separation following extraction of biological samples, may induce the alternation of arsenic species during sample preparation. The present study was aimed to develop novel arsenic speciation methods for biological matrices using surface-enhanced Raman spectroscopy (SERS), which, as a rapid and non-destructive photon scattering technique. The use of silver nanoparticles with different surface coating molecules as SERS substrates permits the measurement of four common arsenicals, including arsenite (AsIII), arsenate (AsV), monomethylarsonic acid (MMAV) and dimethylarsinic acid (DMAV). This speciation was successfully carried out using positively charged nanoparticles, and simultaneous detection of arsenicals was achieved. Secondly, arsenic speciation using coffee ring effect-based separation and SERS detection was explored on a silver nanofilm (AgNF), which was prepared by close packing of silver nanoparticles (AgNPs) on a glass substrate surface. Although arsenic separation using the conventional coffee ring effect is difficult because of the limited migration distance, a halo coffee ring was successfully developed through addition of surfactants, and was shown to be capable of arsenicals separation. The surfactants introduced in the sample solution reduce the surface tension of the droplet and generate strong capillary action. Consequently, solvent in the droplet migrated into the peripheral regions and the solvated arsenicals to migrated varying distances due to their differential affinity to AgNF, resulting in a separation of arsenicals in the peripheral region of the coffee ring. Finaly, a method combining experimental Raman spectra measurements and theoretical Raman spectra simulations was developed and employed to obtain Raman spectra of important and emerging arsenic metabolites. These arsenicals include monomethylarsonous acid (MMAIII), dimethylarsinous acid (DMAIII), dimethylmonothioarinic acid (DMMTAV), dimethyldithioarsinic acid (DMDTAV), S-(Dimethylarsenic) cysteine (DMAIIICys) and dimethylarsinous glutathione (DMAIIIGS). The fingerprint vibrational frequencies obtained here for various arsenicals, some of which have not reported previously, provide valuable information for future SERS detection of arsenicals.