Pulmonary rehabilitation (PR) in patients with COPD improves quality of life, dyspnea, and exercise tolerance. However, 30 to 50% of patients are “non-responders” (NRs) according to considered variables. Surprisingly, peripheral muscle force is never taken into account to attest the efficacy of PR, despite its major importance. Thus, we aimed to estimate the prevalence of force in NRs, their characteristics, and predictors of non-response. In total, 62 COPD patients were included in this retrospective study (May 2019 to December 2020). They underwent inpatient PR, and their quadriceps isometric maximal force (QMVC) was assessed. The PR program followed international guidelines. Patients with a QMVC increase <7.5 N·m were classified as an NR. COPD patients showed a mean improvement in QMVC after PR (10.08 ± 12.97 N·m; p < 0.001). However, 50% of patients were NRs. NRs had lower pre-PR values for body mass, height, body mass index, PaO2, and QMVC. Non-response can be predicted by low QMVC, high PaCO2, and gender (when male). This model has a sensitivity of 74% and specificity of 81%. The study highlights the considerable number of NRs and potential risk factors for non-response. To systematize the effects, it may be interesting to implement blood gas correction and/or optimize the programs to enhance peripheral and central effects.