A new method of measuring the albedo of urban prototype is proposed. The method is used to measure ten urban prototypes with different pavement reflectivity and with south-north orientation, west-east orientation and cross-street orientation, respectively. The results are compared with those obtained by the ASTM E1918-06 and the modified ASTM E1918-06. It is found that when the variation of the incident solar intensity is less than 20 W/m²(a tolerant error stated by ASTM E1918A), the ASTM E1918-06 can either underestimate or overestimate the albedo of the urban canyon prototype up to 0.10. For an urban canyon(UC) with an aspect ratio of 1.0, an change from 0.15 to 0.65 of pavement albedo would cause an increase of the albedo of the UC from about 0.15 to 0.35 if the albedo of the roof and wall is about 0.40. Raising the albedo of the pavement in a UC is not an effective way to increase the albedo of the urban area, especially for UC with great aspect ratio. For low aspect ratio UC, raising the albedo of the pavement or of the parking lot introduces a sizable additional diffuse reflected radiation to the pedestrians. Therefore, it should be cautious to developing reflective pavements as an urban cooling strategy.