Abstract: As a photochemical produced oxidant, gaseous hydrogen peroxide (H₂O₂) plays an important role in aerosols and acid rain production. However, not many measurements of gaseous H₂O₂ have been made in China in the past few years and further studies especially the level and trend of H₂O₂ in smog are needed. To make up for the lack of H₂O₂ data and provide support for air quality improvement in China, an observation experiment on gaseous H₂O₂ is carried out from 27 Dec 2016 to 28 Apr 2017 at an urban site in the northwest of mega-city Beijing, using a two-channel H₂O₂ monitor AL2021. The concentration level, variation and influence factors under different conditions are analyzed with several simultaneously observed pollutants (O₃, PAN, NO_X, PM_2.5, etc.) and meteorological parameters. The mean mixing ratio of H₂O₂ for the entire period is (0.65±0.59)×10^-9, with a higher mean of (0.83±0.67)×10^-9 in spring and a lower mean of (0.51±0.47)×10^-9 in winter. Day peaks with the value higher than 2×10^-9 is also detected in winter indicating that high concentration of H₂O₂ can also happen under certain conditions. The concentration of H₂O₂ shows pronounced diurnal cycles with peaks in the period of 1800-2100 BT, occurring later than those reported for other sites in China or foreign countries and shows a delay of about 4 to 7 hours compared with the peaking time of O₃. H₂O₂ level is found to be negatively correlated with relative humidity (RH), especially when only considering the maximum H₂O₂ level under RH over 55%. This is consistent with the uptake of gaseous H₂O₂ by water-containing aerosol particles under higher RH conditions. The H₂O₂ peaking time and peak level are closely related with RH as well as other factors, such as NO_X. Under conditions of daily RH lower than 55%, H₂O₂ level can reach a mean peak value of 1.52×10^-9 with peaking times during 1800-2400 BT, while it peaks lower (1×10^-9) and earlier (before 1700 BT) with the daily mean RH higher than 65%. H₂O₂, O₃ and PAN show different diurnal patterns and levels under different pollution conditions. H₂O₂ shows smaller average level differences for clean and hazy days, with a higher peak but a lower level during 1100-1500 BT under the clean condition. O₃ shows a higher mixing ratio under clean condition than under hazy condition, while PAN reveals an opposite trend. Results also indicate that dynamical transport could be an important influencing factor of variations and levels of H₂O₂ and O₃. The impact of photochemistry on haze formation in colder months in the urban environment of Beijing and its feedback warrant further studies, particularly the role of H₂O₂ in the formation of sulfate aerosol.