Aiming at the challenges of difference between the research conclusion and actual situation resulting from inadequate selection of research scale and incomplete research dimension in the study of the urban thermal environment, an approach of quantitative analysis for the correlation between urban building spatial distribution pattern and urban thermal environment is suggested. First, from Landsat8 remote sensing images, the land surface temperature (LST) in the research area is extracted. Then, the quantitative expression system of urban building spatial distribution pattern in the research area is built. Next, correlation analysis (CA) and bivariate local spatial autocorrelation analysis (BLSAA) are employed to investigate the effect of urban building spatial distribution patterns on the thermal environment at 1- and 2-km scales. The investigation findings of the suggested approach to the thermal environment of the central urban area of Jinan demonstrate that one-, two-, and three-dimensional urban spatial structure indicators are sensitive to the change in research scale; one- and three-dimensional urban spatial structure indicators are extremely sensitive. The confidence of investigation findings on the 1-km scale is higher than that on the 2-km scale. The one-, two-, and three-dimensional urban spatial structure indicators' effect on the urban thermal environment is a primarily positive correlation, and the number of positive correlation indicators increases with the increase in dimension. The two- and three-dimensional urban spatial structure indicators have a greater correlation with the urban thermal environment compared with the one-dimensional urban spatial structure indicator.