Occupant-related uncertainty has been recognized as one of the main challenges that building designers face. Current occupant modelling practices are based on simple assumptions that are typically made based on codes, standards, and rules-of-thumb. Designers assume occupants have homogeneous temporal and spatial distribution. This approach does not recognize differences among tenants or buildings and can lead to suboptimal design solutions that can compromise energy and comfort performance. Therefore, this doctoral research aims at developing a practical improved method for occupant modelling that recognizes occupant-related uncertainty. The method was developed based on a thorough qualitative and quantitative investigation. To this end, a deeper understanding of the current approaches, challenges and needs of occupant modelling throughout the design process was obtained and documented through a stakeholders' workshop and interviews with a case study design stakeholders. Then, a simulation-based investigation was conducted on a real case study office building located in Toronto, Canada. The simulation-based investigation included conducting a parametric analysis under variable occupant scenarios, developing an occupant-centric design optimization method, and evaluating the impact of occupants' spatial distribution on energy and comfort performance. The documentation of current occupant modelling practice indicated the need to improve the current approach by carefully considering occupant-related assumptions in early design stages. In addition, it indicated the need to improve communicating occupant-related assumptions among design stakeholders. The results of the simulation-based investigation indicated that occupant-related assumptions can influence the outcomes of design parametric analysis and design optimization. Notably, assumptions about occupants' spatial distributions demonstrated substantial impact on occupants' thermal comfort and the indoor air quality.