A model-based framework is described that permits the optimal composition of cellulase enzyme mixtures to be found for lignocellulose hydrolysis. The rates of hydrolysis are shown to be dependent on the nature of the substrate. For bacterial microcrystalline cellulose (BMCC) hydrolyzed by a ternary cellulase mixture of EG2, CBHI, and CBHII, the optimal predicted mixture was 1:0:1 EG2:CBHI:CBHII at 24 h and 1:1:0 at 72 h, at loadings of 10 mg enzyme per g substrate. The model was validated with measurements of soluble cello-oligosaccharide production from BMCC during both single enzyme and mixed enzyme hydrolysis. Three-dimensional diagrams illustrating cellulose conversion were developed for mixtures of EG2, CBHI, CBHII acting on BMCC and predicted for other substrates with a range of substrate properties. Model predictions agreed well with experimental values of conversion after 24 h for a variety of enzyme mixtures. The predicted mixture performances for substrates with varying properties demonstrated the effects of initial degree of polymerization (DP) and surface area on the performance of cellulase mixtures. For substrates with a higher initial DP, endoglucanase enzymes accounted for a larger fraction of the optimal mixture. Substrates with low surface areas showed significantly reduced hydrolysis rates regardless of mixture composition. These insights, along with the quantitative predictions, demonstrate the utility of this model-based framework for optimizing cellulase mixtures. Biotechnol. Bioeng. 2011; 108: 2561-2570. (C) 2011 Wiley Periodicals, Inc.