Rfn (formaldehyde/nitrogen dioxide) is a common indicator based on satellite observations used to classify ozone formation sensitivity. However, it may underestimate anthropogenic volatile organic compounds (VOCs) in heavily polluted cities when only formaldehyde (HCHO) is used in Rfn to measure VOCs, since it is mainly derived from natural sources worldwide. In this study, we used multiaxis differential optical absorption spectroscopy to acquire tropospheric observations of nitrogen dioxide (NO2), HCHO and glyoxal (CHOCHO) in Beijing from 1 April 2019 to 31 March 2020. Combined with VOCs detected simultaneously by gas chromatography—mass spectrometry and proton transfer reaction–time-of-flight/mass spectrometry near the ground, we evaluated the representativeness of HCHO column densities on total VOCs (TVOC) in equivalent propylene concentrations, which is called reactivity. The results showed that there were significant seasonal differences in the response of HCHO to TVOC reactivity, with fitting slopes of 2.3 (spring), 2.6 (summer), 2.9 (autumn) and 1.0 (winter) in the four seasons, respectively. Since CHOCHO can be used to partly characterize the contribution of anthropogenic VOC emissions and demonstrated a better response to TVOC reactivity in winter, with fitting slopes of 0.2 (spring), 0.2 (summer), 0.2 (autumn) and 0.5 (winter) in the four seasons, respectively, we introduced CHOCHO to construct a new indicator (HCHO + 6 × CHOCHO). The fitting slopes of the four seasons were more similar, being 3.2 (spring), 3.6 (summer), 4.0 (autumn) and 4.0 (winter). The ratio of the new indicator to NO2, Rmn ((HCHO + 6 × CHOCHO)/NO2), was used to reclassify the ozone formation sensitivity of urban areas in North China, revealing that it is a transition regime before 1300 LST (LST = UST + 8) and an NOx-limited regime afterwards. Rmn improved the sensitivity from the VOC-limited regime to the NOx-limited regime, enhancing the sensitivity of NOx and providing new robust support for ozone pollution prevention and control.