Abstract: To address the challenges of strong reservoir heterogeneity in coalbed methane (CBM) horizontal wells and the reliance on empirical experience with difficulty in quantitative evaluation for conventional fracturing stage design, an intelligent fracturing stage optimization method based on multi-parameter fusion was proposed. Coal seam encounter rate, total hydrocarbon content, gamma ray value, and drilling time were selected as the main evaluation parameters. Positive and negative normalization was employed to construct a comprehensive evaluation index, and constraints including intra-stage quality, intra-stage homogeneity, and inter-stage heterogeneity were introduced to establish an intelligent fracturing stage optimization model. Dynamic time warping (DTW) was used to identify local morphological similarities among logging curves, and the particle swarm optimization (PSO) algorithm was applied for global optimization of stage boundaries, forming a hybrid DTW-PSO solution workflow. The stage design results of Well PS-4-2 in Huainan demonstrated that the optimized stage corresponded well to the characteristics of coal seam development, gas-bearing properties, and lithological variations along the horizontal section. The coal seam encounter rate and total hydrocarbon content contributed the most to the comprehensive score, with Stage 1 and Stages 5–10 identified as favorable stages, Stages 2 and 11 as secondary favorable stages, and Stages 3–4 as unfavorable stages. After field fracturing of Well PS-4-2, the gas production continued to rise, while the water production gradually decreased and stabilized, which indicated a significant fracturing stimulation effect and validated the engineering applicability and effectiveness of the proposed method. The research results can provide technical support for refined fracturing design and efficient development of CBM horizontal wells.