In this study, we propose a method that combines feature point matching and measurements information for similarity search in 2D-CAD drawings. Conventional feature point matching methods often overlook measurements information
, leading to high similarity scores for parts with similar shapes but different measurements. Our method incorporates numerical measurements information extracted from drawings as a correction factor in the similarity calculation. This allows the method to assign lower scores to parts with similar shapes but differing measurements. Evaluations using multiple types of part drawings showed that considering both shape and measurements information improves similarity search accuracy.