Abstract: To address the strong reservoir heterogeneity of coalbed methane (CBM) horizontal wells and the limitations of conventional fracturing stage design, which largely relies on empirical judgment and lacks quantitative evaluation, this study proposes an intelligent stage optimization method for hydraulic fracturing based on multi-parameter integration. Coal seam encounter rate, total hydrocarbon content, gamma ray value, and rate of penetration were selected as the main evaluation parameters. A comprehensive evaluation index was established through positive and negative normalization. Intra-stage quality, intra-stage homogeneity, and inter-stage heterogeneity were introduced as constraints to construct an intelligent optimization model for fracturing stage division. Dynamic time warping (DTW) was used to identify local morphological similarities among logging curves, while Particle swarm optimization (PSO) was applied to globally optimize stage boundaries, forming a DTW-PSO hybrid solution workflow. The method was applied to Well PS-4-2 in Huainan. The results show that the optimized stage division corresponds well to the variations in coal seam development, gas-bearing properties, and lithological characteristics along the horizontal section. Coal seam encounter rate and total hydrocarbon content contribute most significantly to the comprehensive score. Stage 1 and stages 5–10 are identified as favorable intervals, stages 2 and 11 as moderately favorable intervals, and stages 3–4 as unfavorable intervals. After field fracturing stimulation, gas production increased continuously, while water production gradually declined and stabilized, indicating a significant stimulation effect and verifying the engineering applicability and effectiveness of the proposed method. This study provides technical support for refined fracturing design and efficient development of CBM horizontal wells.