tBuilding thermal insulation (BTI) may influence indoor cooling by green roofs. The hypothesis that build-ings with poor BTI may benefit more from green-roof passive cooling is tested. An experiment wasdesigned in humid-tropical Hong Kong to investigate two BTI states (good and poor) and three treat-ment plots (control-bare, succulent Mexican Sedum, and broadleaved Perennial Peanut). Temperaturesensors were installed along a holistic vertical profile from outdoor air to green-roof layers and indoorceiling and air. On summer-sunny day, poorly-insulated control roof allows notable daytime peak thermalintrusion, counterpoised by comparable nocturnal thermal discharge. It is the only roof with bidirectionalheat fluxes and nearly-balanced diurnal heat-energy budget. Other roofs are predominantly unidirecti-onal (downward) with considerable indoor heat gain. Well-insulated control roof experiences extended(morning to night) thermal intrusion, with depressed but delayed and prolonged heat influx, and thermalinsulation breaching. At night time, it creates nocturnal thermal barrier to reduce heat escape from indoorspace. Simple Sedum green roof on poorly-insulated roof establishes green-roof heat-sink effect to incurdaytime and nocturnal aggravated thermal intrusion. On well-insulated roof, it generates a synergistic-dual thermal barrier to suppress thermal insulation breaching. However, nocturnal thermal discharge hasbeen constrained by persistent positive thermal gradient throughout the day. Thicker and more-elaboratePeanut green roof brings considerable evapotranspiration cooling and synergistic-dual thermal barrierto generate daytime subdued thermal intrusion and mask inherent BTI-differences. As omission of BTIfor both Sedum and Peanut green roofs does not earn more indoor cooling, the hypothesis cannot beaccepted. To optimize passive cooling, green roofs with thicker substrate and denser foliage should beinstalled on buildings with good BTI.