Positive pressure mechanical ventilation (MV) has been utilised in the care of critically ill patients for over 50 years. MV essentially provides for oxygen delivery and carbon dioxide removal by the lungs in patient with respiratory failure or insufficiency from any cause. However, MV can be injurious to the lungs, particularly when high tidal pressures or volumes are used in the management of Acute Respiratory Distress Syndrome (ARDS) or similar acute lung injuries. The hallmark of ARDS is extensive alveolar collapse resulting in hypoxemia and carbon dioxide retention. Application of Positive End Expiratory Pressure (PEEP) is used to prevent derecruitment of alveolar units. Hence, there is a delicate trade-off between applied pressure and volume and benefit of lung recruitment. Current clinical practice lacks a practical method to easily determine the patient specific condition at the bedside without excessive extra tests and intervention. Hence, individual patient treatment is primarily a mixture of "one size- fits-all" protocols and/or the clinician's intuition and experience. A quasi-static, minimal model of lung mechanics is developed based on fundamental lung physiology and mechanics. The model consists of different components that represent a particular mechanism of the lung physiology, and the total lung mechanics are derived by combining them in a physiologically relevant and logical manner. Three system models are developed with varying levels of physiological detail and clinical practicality. The final system model is designed to be directly relevant in current ICU practice using readily available non-invasive data. The model is validated against a physiologically accurate mechanical simulator and clinical data, with both approaches producing clinically significant results. Initial validation using mechanical simulator data showed the model's versatility and ability to capture all physiologically relevant mechanics. Validation using clinical data showed its practicality as a clinical tool, its robustness to noise and/or unmodelled mechanics, and its ability to capture patient specific responses to change in therapy. The model's capability as a predictive clinical tool was assessed with an average prediction error of less than 9% and well within clinical significance. Furthermore, the system model identified parameters that directly indicate and track patient condition, as well as their responsiveness to the treatment, which is a unique and potentially valuable clinical result. Full clinical validation is required, however the model shows significant potential to be fully adopted as a part of standard ventilator treatment in critical care.