A randomly oriented nanowire network, also called nanonet (NN), is a nano-microstructure that is easily integrated into devices while retaining the advantages of using nanowires. This combination presents a highly developed surface, which is promising for sensing applications while drastically reducing integration costs compared to single nanowire integration. It now remains to demonstrate its effective sensing in real conditions, its selectivity and its real advantages. With this work, we studied the feasibility of gaseous acetone detection in breath by considering the effect of external parameters, such as humidity and temperature, on the device’s sensitivity. Here the devices were made of ZnO NNs covered by SnO$_2$ and integrated on top of microhotplates for the fine and quick control of sensing temperature with low energy consumption. The prime result is that, after a maturation period of about 15 h, the devices are sensitive to acetone concentration as low as 2 ppm of acetone at 370 °C in an alternating dry and wet (50% of relative humidity) atmosphere, even after 90 h of experiments. While still away from breath humidity conditions, which is around 90% RH, the sensor response observed at 50% RH to 2 ppm of acetone shows promising results, especially since a temperature scan allows for ethanol’s distinguishment.