Alloys are widely used in our lives and their classification is an important part of metal recycling. However, with similar appearance, density, and electromagnetic (EM) properties, alloys are more difficult to categorize in actual metal classification. In this article, a classification method for alloys with high conductivity resolution is proposed based on single-frequency eddy current (EC) sensing technology. Experimentally, eight kinds of alloys with a minimum difference in conductivity of 0.84 MS/m in the range of 16.09 to 29.39 MS/m were classified. We have analytically deduced the proportionality of the proposed feature to the conductivity and provided evidences by the finite element method (FEM). Issues of coil radius design, excitation frequency setting, and lift-off effect suppression are addressed in this article. The proper designs of the coil radius and the excitation frequency of the EC sensor are essential for the improvement of conductivity classification resolution. Lift-off effects of EC sensing due to vibration in industrial conveyorization are also considered, and we realized the lift-off immunity by sensor calibration with the known lift-offs. We innovatively propose to analyze the EC sensor design and optimization issues in view of the coupling of excitation frequency and conductivity. The proposed methodology improves the resolution and rate of EC sensing-based conductivity classification effectively.