Therapist driven protocols may help to tailor rehabilitation programs to individual patients. We aimed to test the feasibility, safety, and clinical usefulness of a therapist driven protocol for rehabilitation including exercise training of patients with heart or lung diseases. An internal audit elaborated the Cardio-Respiratory Exercise Maugeri Algorithm (CREMA) based on: (a) standardized baseline assessments, (b) decision-making pathways, and (c) frequency/intensity/time/type (FITT) of prescription for each exercise. Outpatients (n = 620) with chronic heart disease (CHD), recent myocardial revascularization (REVASC), chronic airway (Obstructive), and restrictive lung (Restrictive) diseases underwent exercise training according to CREMA during 4 years. Peripheral muscle strengthening was the most prescribed exercise (83.6%), while arm endurance training was the least frequently (0.75%). Exercise prescription varied widely among the disease groups (interval training 19–47%, balance 35–49%, lower limb muscle training 6–15%). After training, REVASC patients were the best improvers in the 6 min walking distance (+48.7 (56.1) m), maximal inspiratory pressure (+9.6 (15.4) cmH2O), and daily steps (+1087.2 (3297.1) n/day). Quadriceps and biceps strength, maximal expiratory pressure, and balance improved in all groups, without significant differences. Minor side effects were observed in 11.2% of the patients. The CREMA therapist driven protocol was feasible, safe, and useful for prescribing tailored training programs. Exercise prescriptions and training response differed among diseases.