The transition to circular textile manufacturing necessitates the effective incorporation of recycled fibers while ensuring consistent fabric performance. This research demonstrates the feasibility of producing circular knitted fabrics from recycled cotton, eco-viscose, and modal tri-blend fibres under industrial knitting conditions using full-scale machinery, and subsequently introduces a comparative modeling framework to evaluate the effects of recycled cotton content, yarn count, and stitch length on fabric areal density and bursting strength. A Box-Behnken design was employed to generate knitted fabric samples under controlled processing conditions. Quadratic regression and fuzzy linear regression were used to model the relationships between process variables and fabric responses. Both models demonstrated strong predictive capability, indicating that recycled fibre content interacts significantly with yarn count and stitch length in determining fabric compactness and mechanical resistance. Increasing yarn count and optimizing stitch length were found to enhance fabric structural stability. This study integrates predictive modeling with circular material utilization to establish a framework that enhances processes and forecasts performance in the production of fabric derived from recycled cotton fibers. The resulting fabrics showed GSM values of 118-160 g/m² and bursting strength of 549-853 kPa, indicating stable fabric structure and adequate mechanical performance even with recycled cotton incorporation.

circular economy, recycled fibre, fuzzy linear regression, quadratic regression, Box-Behnken design