Abstract In wind-resistant designs of high-rise buildings, wind profiles are modeled for strong wind, where atmospheric stability is assumed to be neutral. When wind profiles are modeled from the results of short-term observations, which include few data on strong wind, a method to evaluate atmospheric stability and detect wind profiles under neutral conditions is required. In the present study, the applicability of the Pasquill stability class, a widely used index to evaluate atmospheric stability, was assessed in urban areas. The atmospheric stability evaluated from the Pasquill stability class was compared with that from the Monin–Obukhov length using the observation data obtained in Tokyo. 70% of the data of Pasquill stability class D (neutral class) corresponded to the neutral condition evaluated from the Monin–Obukhov length. Furthermore, the mean wind profiles, which were observed at another site in Tokyo using Doppler lidar, were classified according to the Pasquill stability class. As the Pasquill stability class transformed from unstable to neutral, the power law exponent increased. For the neutral class, the power law exponent was within the error range of previous observations of strong wind. The mean wind profile of the neutral class can approximate that of strong wind appropriately.