This paper shows a comparison between predicted cooling conditions, site measurements and user's responses into two type of buildings designed to be naturally cooled under tropical conditions. The monitored buildings are two commercial malls and a large industrial building located in Colombia. Regarding predicted cooling conditions, adequate strategies were addressed to restore comfort: Air movement, solar direct gain control and thermal mass control were encouraged in a balanced way. Some devices were introduced to assure bioclimatic requirements as ventilated sky lights, transparent solar control polycarbonate or reinforced polyester roofing and double skin or low –E thermal sheets walls. After the construction process was finished, efficiency tests was carried on into the selected buildings. According to the predicted requirements some of them were allowed, specially the internal temperature ranges and air changes in order to exhaust overheating and hygienic air removal. Nevertheless, internal air speeds were variable in a large range depending on exterior wind, which are also variable under the Colombian conditions. Under occupied conditions, users were asked to know their comfort perceptions. After setting a crossway analysis, relevant facts were founded: 1. The comfort effect of convection air movement produces a differential temperature of 3oC -4oC between different commercial mall levels. This has a direct impact on user comfort behaviour. 2. Some air movement, at minimum speeds (0.10 m/sec- 0.30 m/sec) produced an adequate user response, even if predicted air movement was established to produce 0.50 m/s. 3. Regarding solar and thermal mass devices, users response suggest a satisfaction perception compared with external maximum temperatures over 30 oC. (not exceeded 38oC) 4. Into the industrial building, due to a high thermal stratification, minimum solar gains trough envelope and continuous overheat exhaust by ventilated skylights, the worker responses were acceptable, corresponding to site measurements and predicted requirements.