Natural ventilation in both urban and interior space has great influence on air quality, thermal comfort and energy consumption. Hong Kong is facing serious ventilation challenges at both urban and indoor levels. With limited land area and a growing population, urban regeneration in the old urban area has been and will be an important way forward for Hong Kong in the future. Urban fabric initially planned for medium-rise buildings will gradually be replaced by high-rise high-density buildings. Both urban ventilation and indoor natural ventilation performance will be affected due to this morphological change and therefore need to be investigated. After careful validation, CFD simulation using the k-ω model (SST) was performed to study the urban and interior spaces regarding natural ventilation performance on various residential building clusters and unit types. A general survey of historical housing development from the pre-colonial period in Hong Kong as well as a comprehensive survey of residential buildings in the recent 10 years was carried out. A systematic architectural analysis of the urban context and building types in Hong Kong considering factors such as street grid, site type, site area, building typology, building height, plan shape, unit type, unit area, opening configuration, floor area ratio, units per floor and viewing direction were all analyzed to generate realistic cases for ventilation study in the CFD simulation. First, five different street block division strategies with several variations in tower type and podium geometry were created based on the systematical architectural analysis. The urban ventilation performance of each urban scenario was quantified using CFD simulation with SSW, SW, WSW and W incoming wind by indices such as age of air and velocity ratio. The result indicates that street-level ventilation can be improved by increasing the podium permeability or adopting a step-shaped section or aligning the towers to the leeward side of the podium. Street block division strategies with areas smaller than 500 sq.m. such as S1 in the case study should be avoided due to its poor urban ventilation performance. Division strategies with larger site areas such as S5u2 had the best urban ventilation performance, while it also had the highest pollutant concentration at the façade of the tower which indicates potential indoor natural ventilation problems. Then, indoor natural ventilation of five groups of unit types (36 variations) and inter-unit pollutant transmission between the units were tested using CFD simulation. Unit “a4” had the best ventilation performance among type “a” units and “b3-c” had the most balanced performance in both indoor and urban ventilation among type “b” units. Better ventilation performance in type “c” units than types “a1” and “a2” indicates that arranging openings on both sides of the façade is more effective than the provision of a deep re-entrant. Unit “d2” was the best ventilated unit among unit group “d” and indicates that openings in different rooms should not be aligned. Single-side ventilation of small unit types was found to be better when the units were placed in a smaller plan than in a larger plan. Larger re-entrant width, open kitchen plan and suitable arrangement of openings, casement and side wall were all proved to have positive influences on indoor natural ventilation. ; published_or_final_version ; Architecture ; Doctoral ; Doctor of Philosophy