Thermal radiation attenuation is a well-known positive effect of water sprays applied to firefighting. It was studied here to quantify its real effect on the whole process of fire mitigation. For the demonstration, three different real scale spraying devices were used: a water mist nozzle operating at 85 bars, a sprinkler operating at low pressure (0.3 bars at the nozzle) and a sprinkler operating at a higher pressure (1.3 bars at the nozzle). The radiation attenuation was specifically studied, separately from other actions of the sprays on the fire, using an experimental configuration where water was injected between the radiation source and a detector, but not directly onto the actual fire. Hence, the radiation flux reduction could only be attributed to mixing of water droplets and smoke, not to any reduction in heat release due to water directly cooling the fire area nor to inerting effects occurring simultaneously. The flux was measured using an infrared device combining a Fourier transform infrared spectrometer and a multispectral infrared camera. Two radiative sources were used: a high-temperature blackbody for quantification of the radiation transmission with a calibrated source, and a 200 kW heptane pool fire for an application to a realistic radiation source involved in fires. When using a blackbody source for radiation generation, the radiation attenuation of the incident fluxes was 89% for the high-pressure water mist, 62% for the sprinkler at the higher operating pressure and 20% for the sprinkler at the lower pressure. These differences were easily explained by a sensitivity study conducted coupling the Mie theory for the prediction of droplet radiative properties to the Monte Carlo method for radiative transfer simulation. This showed the sensitivity of the radiation attenuation to droplet size and concentration. When the blackbody source was replaced by a pool fire, the flux reduction was even higher for the low pressure sprinkler, i.e. 45%, while being almost unchanged for the water mist. In the present configuration, where smoke stratification was observed before spray activation, this was mainly attributed to the fact that a mixing of droplets and smoke was produced in the measurement area due to the drag effect. The smoke and spray mixing resulted in a stronger attenuation capacity with the sprinkler device. This increased attenuation was not observed through the transmission data for the water mist, perhaps because of compensating effects related to evaporation and the surrounding temperature, or a smoke flow that was highly penalized by the spray activation.