A chemiresistive sensor array was created from single-walled carbon nanotubes (SWCNTs) by non-covalent modification with late first-row transition metal complexes of meso-tetraphenylporphyrin. The responses to vapors of various volatile organic compounds (VOCs) were strong and were subjected to statistical analyses that enabled the successful classification of representative VOCs into five different categories (aliphatic hydrocarbons, alcohols, ketones, aromatic hydrocarbons, and amines) with 98% accuracy. With the exception of amines, which are capable of strong charge transfer interactions, the basis of classification appears to correlate with the differences in the solubility properties of the porphyrin compounds in the various VOCs as solvents. This feature suggests that an analyte with greater intermolecular affinity for the SWCNT-porphyrin composite will induce a greater response. These results further demonstrate the potential for porphyrin-functionalized SWCNT-based electronic noses for applications in inexpensive, portable chemical sensors for the identification of VOCs.