Purpose <p>Disrupted cellular redox balance is associated with various diseases, including osteoarthritis. Although mitochondrial (MT) membrane potential is a proxy for redox balance, the translational potential of this method is limited by exogenous dye. Therefore, the objective of this study was to predict changes in MT membrane potential in response to mechanical loading using dye-free optical redox imaging (ORI). A secondary objective was to determine the effect of loading on ORI metrics.</p> Methods <p>Full-thickness porcine cartilage strips were subjected to tensile loading at one of two strain rates (1.00&#xa0;s⁻¹ or 0.10&#xa0;s⁻¹). ORI was done before, immediately after, and 30 minutes after loading. MT membrane potential was then measured using fluorescent dye. A generalized linear mixed-effects model (GLMM) tested main effects (ORI metrics, loading vs. control, loading rate, post-loading time, zone) and their interactions in prediction of MT membrane potential. Significant predictors were retained in a new GLMM that was trained using 70% of the dataset and evaluated using the remaining 30%. Two separate GLMMs evaluated the main effects on ORI metrics.</p> Results <p>In the GLMM using MT red/green ratio as the dependent variable, ORI metrics, loading rate, and loading vs. control were significant main effect. GLMMs to predict MT red/green from ORI that retained significant main effects resulted in an average difference between predicted and actual values of 7.07%. When analyzing the effect of loading vs. control, loading rate, and zone on ORI metrics, only loading rate showed significance.</p> Conclusion <p>ORI can predict MT membrane potential measured by fluorescent dye and has the possibility to be developed as a clinical tool to evaluate cartilage redox balance.</p>