© 2019 Elsevier Ltd HO2 radicals play important roles in tropospheric chemistry. The large discrepancies among field measurements and sophisticated model predictions for the overall HO2 concentrations can be attributed to the HO2 uptake coefficients onto ambient aerosols (γ) have not yet been properly quantified. This study presents the first online measurement of the total HO2 reactivity caused by the ambient gas phase (kg') and aerosol phase (ka') in the summer of 2018 in Kyoto, Japan, using a combined laser-flash photolysis and laser-induced fluorescence (LFP–LIF) technique coupled with a versatile aerosol concentration enrichment system (VACES) that enriches ambient aerosols. The results denote that kg' ranged from 0.1 s−1 (25th percentile) to 0.32 s−1 (75th percentile) with an average value of 0.22 ± 0.16 s−1 (1σ), which can be primarily explained by the reaction of HO2 with NO2. With the application of VACES and the auto-switching aerosol filter, ka' ranged from 0.004 s−1 (25th percentile) to 0.028 s−1 (75th percentile) with an average value of 0.017 ± 0.015 s−1. When converted ka' to ambient conditions (by dividing with the enrichment factor), this result was ~10 times higher than the HO2 reactivity caused by its self-reaction under ambient concentration levels (~5 ppt) at 298 K. The related γ ranged from 0.08 (25th percentile) to 0.36 (75th percentile), with an average value of 0.24, which is comparable with the values used in previous modeling studies (~0.2); however, it presents a large variation of ±0.20 (1σ) within the measurement time. This indicates that a large bias may exist with respect to the estimation of HO2 concentrations when using a constant γ value. Further, we suggest that different γ values should be applied in modeling studies depending on the environment conditions; the γ value obtained here can be used as an upper limit value for urban areas in the summer due to possible aerosol phase changes. The analysis of ambient air backward trajectories indicates that the predominant NO2 emission sources originated from the mainland of Japan. However, no significant differences regarding HO2 uptake coefficients were observed when air masses came through the mainland or from the coastal direction. This study provides considerably reliable γ, which could increase the modeling accuracy of heterogeneous processes in tropospheric chemistry. However, combining online and offline methods appears to be more suitable for addressing the variation in γ and model discrepancies in further studies.