The energy yield of solar thermal installations is influenced by local climatic conditions, heating demand, and the efficiency of the system. These factors are affected by uncertainties arising from natural variations and a lack of knowledge. Designers typically address these uncertainties by oversizing the collector area. However, this practice usually leads to high costs, low efficiency, and short service lifetimes of the system. This paper proposes a robust sizing methodology for the solar thermal system used for service hot water preparation in a case-study hospital with a daily average hot water consumption of 8.69 m3. This novel approach incorporates uncertainties through probabilistic simulations, assessing the system’s reliability in achieving a 70% solar fraction design goal. The simulations reveal a mere 22% confidence in the initial deterministic design concerning the solar target and emphasize the importance of thermal insulation, temperature control, and collector cleaning as key factors to reduce performance variability. Although the final revised design outcome requires a 15% increase in the collector area compared to the initial scenario and stricter specifications for control components, pipe insulation, and valve maintenance, it achieves a 90% reliability. In addition, annual carbon emissions and levelized costs of hot water production can be reduced by 18% and 13%, respectively.