Quantifying the relationship between land use change and landscape pattern (LP) is crucial in responding global challenges such as climate change, food security, and even biodiversity loss. Here, we used multi-period land data to quantify the spatiotemporal trends in built-up land expansion (BLE) and LP in the Yellow River Basin (YRB) during 1980–2018. We fitted the correlations between these two variables using LP Index, Landscape Expansion Indices (LEI) and the Multiscale Geographically Weighted Regression (MGWR) model. We found that regional BLE was highly variable, BLE was the main driver of landscape fragmentation, and the relationship between BLE and LP scale varies, especially in terms of the edge BLE. Landscape fragmentation and aggregation spatially varies and temporally increases in complexity. The MGWR model effectively disclosed a scale effect between BLE and LP. The control variables differed in terms of their relative impact on LP evolution. The present study demonstrated the complexity of the relationship between BLE and LP change and highlighted the importance of variable selection, spatiotemporal scale, and spatial dimension in the quantification of these relationships.