Vegetation landscaping around the Metro overpass can impact the noise attenuation effect in surrounding residential areas. Six residential buildings with vegetation landscaping situated along the elevated Metro overpass of Hangzhou Metro Line 9 were chosen to assess their noise levels on sites at the 1st to 5th floors. Quantitative investigation of the vegetation landscaping condition and the noise loudness was conducted in-situ, and the greening conditions of the vegetation were investigated quantitatively, including canopy closure, greening rate, height under the lowest branch of the arbour, number of arbours, and the effective spatial occlusion coefficient (ESOC), etc. The in-situ noise loudness of each measurement site was also determined. A 3-D model of community was also created in Rhinoceros software and imported into Cadna/A software to calculate the simulated noise loudness of the scenario without vegetation. The net noise attenuation coefficient by vegetation (NACV) was calculated to quantify the noise reduction effect by vegetation landscaping. A Python statistical program is applied to perform a linear regression analysis and propose a mathematical model for describing noise attenuation effect. It indicated that the ESOC, canopy closure and the greening rate could jointly explain 71.6% of variations in mean NACV. The influence of factors was found in the following order: distance between the vegetation and the noise source > height of the noise source > greening rate > ESOC. This study reveals the characteristic factors and correlations of noise attenuation and explores the effect of vegetation landscaping on urban noise attenuation.