TSU (territorial space use) constitutes the foundational behavior of human economic and social activities upon the land. Investigating the threshold effect of multi-scale TSU on LES (land ecological security) is essential for scientifically constructing a border ecological barrier management system within the framework of integrated development and security planning. Currently, spatial governance of territorial space in China's border areas utilizes the "three districts and three lines" as its core framework, implementing rigid spatial controls through the ecological protection red line, cultivated land and permanent basic farmland boundaries, and the urban development boundary. Concurrently, the main functional area strategy is employed to promote differentiated development. However, these border areas face dual pressures stemming from ecological fragility and intensive human activity: the northern border is threatened by ecological degradation in arid regions; soil erosion affects more than half of the northwest border area; and rocky desertification in the southwest border region contributes to the desertification of cultivated land and the loss of soil fertility. Consequently, a significant spatial mismatch and structural imbalance exist between land use patterns and ecological security requirements. By employing a BRT (Boosted Regression Trees) model and SEM (Structural Equation Model), this study identified and categorized the influence of major TSU types on LES. Subsequently, hierarchical governance of LES thresholds was implemented through SLR (Segmented Linear Regression). Concurrently, we employ the SOM-K-means clustering method to execute zonal control of the predominant types of TSU. The results show that (1) from 2008 to 2023, the proportion of OE (Other ecological space) in the types of TSU is the highest at 32.764 %, which is concentrated in the Northern border and the Northwestern border; OE -> WE (Water ecological space) reflects the systematic tendency of transformation; and the spatial transfer of the three regions is mainly dominated by Agriculture -> Urban and Ecology -> Urban. (2) From 2008 to 2023, LES at different scales is characterized by 'high in the North and low in the West,' and SP (Social protection), EM (Economic drive), and SS (Ecological support) in each border region show strong effects on LES, with AP (Agricultural production space), GE (Grassland ecological space), OE, and FE (Forest ecological space) thresholds for LES of 0.457, 0.425, 0.330 and 0.345, respectively, 0.330, 0.348. (3) In 2023, the grid scale is mainly dominated by WE, FE, AP et al., and the distribution of its KCA (key conservation areas) is as high as 38.713 %; the county and city scales are mainly composed of FEL (Forest ecological space dominant type), APL(Agricultural production space dominant type), GEL (Grassland ecological space dominant type), and OEL (Other ecological space dominant type) to form the spatial pattern of the territorial space, and the RRA (remediation and restoration areas) of the county have the highest percentage (42.244 %), and the KCA of the city area account for 37.686 %. In light of the integrated governance zoning characterized by 'Zonal control + hierarchical governance,' we propose a nuanced governance approach termed 'grid-county-city.' This approach seeks to offer a foundational theoretical framework and decision-making benchmark for the sustainable utilization of territorial space, as well as for harmonizing security and development in border regions.