Although destabilization of solvent-exposed amide groups in the coil state has been shown to be the main factor involved in helix induction in alcohol denaturation, no theory has been proposed that describes the alcohol concentration dependence of denaturation. This paper presents a theory that can describe transfer free energies of a peptide backbone unit from water to aqueous alcohol solutions of various compositions. This theory is an indispensable step in the quantitative description of the solvent roles in alcohol denaturation. The theory is based on the definition of the molarity-based transfer free energy and is made up of following two contributions: (1) free site transfer, which is shown basically to be the cavity formation term, and (2) solvent exchange coming from competition of hydrogen bonding of solvents to the peptide. The theoretical transfer free energies successfully reproduce, at a low alcohol concentration range, the experimental values for the case of aqueous methanol and ethanol solvents, where excluded-volume effects are shown to play the dominant role. The theory has further been applied to 2,2,2-trifluoroethanol (TFE), for which experimental values of peptide transfer free energy are not available. The theory goes beyond the Tanford scheme by introducing conditional solvation of the peptide unit by side chains and successfully describes the TFE concentration dependence of the peptide transfer free energy, which reflects the dependence of helix induction curve.