Abstract: An automatic lens parameter detection system based on low-coherence interferometry was designed. The system employed dual-wavelength (850 nm and 470 nm) LED light sources to generate interference signals. To address the underdetermined fitting problem of Cauchy dispersion coefficients under dual-wavelength conditions, a segmented fitting method based on group refractive index differences was proposed to estimate higher-order coefficients, thereby enabling precise inversion of phase refractive index and Abbe number from group refractive index. With an FPGA as the core hardware platform, the system integrated functions such as multi-channel data acquisition, lock-in amplification, CORDIC-based phase calculation, and motor control to achieve high-precision non-contact measurements of thickness, refractive index, and Abbe number. The measurement error of the system was less than ±0.01 mm for thickness, less than ±0.005 for refractive index, and the Abbe number measurement error was reduced from the industry-allowed ±5% to within ±2%. All coefficient of variation values were below 0.37%, demonstrating that the system meets the requirements for high-precision and high-stability online inspection in industrial applications.