This paper is the result of comprehensive survey engineering of natural resources.

Objective It is the inherent demand of ecological protection and high-quality development to measure and analyze the spatio-temporal differentiation of urban land use efficiency in the Yellow River basin.

Methods Based on municipal input-output data from 2004-2017, we introduce SFA (Stochastic Frontier Analysis) and Spatial Association Model to measure the urban land use efficiency of 83 cities located in the Yellow River basin.

Results The urban land use efficiency of the Yellow River Basin showed an obvious increasing trend during 2004 to 2017. The efficiency growth rate of cities in the upper reaches of the Yellow River basin is faster than that of cities in the middle and lower reaches. The efficiency was still in the middle and low level across the whole area.It still presented great differences of the urban land-use efficiency across the whole the Yellow River basin, but the leadership of central cities played a tremendous role in the urban agglomeration, and the high-efficiency areas (first and second types) were gradually expanding from "Spots" to "Patches", while the middle and low-efficiency areas gradually reduced. The differences of inter-urban agglomeration and differences among the upper, middle and lower reaches gradually converged. The land use efficiency among cities had positive spatial correlation and the agglomeration level was enhanced year by year. The Lisa (local indicators of spatial association) was characterized by polarization. Not only the High-High agglomeration (H-H type) area were mainly distributed in Shandong Peninsula, and gradually clustered into pieces, but also it appeared in the Central Plain urban agglomeration for the first time. The areas with a Low-High agglomeration (L-H type) pattern were transformed from Weifang, Shandong Province to Bayannur, Inner Mongolia. And the High-Low agglomeration areas(H-L type) were concentrated in Lanzhou. The Low-Low agglomeration (L-L type) areas were distributed in the national-level concentrated contiguous poverty areas such as Liupan Mountain area and Qinba Mountain area.

Conclusions It is suggested that as a whole, the Yellow River basin should be planned and developed in a coordinated way, and more attenion should be paid to the endogenous impetus of urban development individually. The spatial spillover effect which the central cities make-should be given full play to, and resources should be allocated and shared rationally through the optimization of the scale and structure of capital investments. In addition, to better serve the transformation of urban economic development from labor-intensive and capital-dependent to technology-driven, the policy of differentiated ecological protection and high-quality development should be implemented in the upper, middle and lower reaches of the Yellow River Basin.