Ecological barrier of national land space can be constructed to utilize the spatial-temporal evolution and influencing mechanism of landscape ecological risk in Xiuhe River Basin. It is of great significance to optimize the ecological risk management. This study aims to take the typical area with the mixed distribution in plain and mountainous areas in the time points of 2002, 2007, 2012, 2017 and 2022. Firstly, the grid database was constructed to cover the land use, human activities, climate and geographical location using GEE platform and ArcGIS 10.8 platform. Secondly, the landscape ecological risk of each evaluation unit was calculated using the landscape index. The spatial distribution of landscape ecological risk was then obtained by kriging interpolation. In addition, the spatial autocorrelation and topographic gradient classification models were used to reveal the spatial and temporal evolution of landscape ecological risk in horizontal and vertical space. Finally, the optimal parameter geographic detector was used to select the indicators from the human activities, climate conditions, natural environment and dimensions. A summary was made to clarify the influence and interaction mechanism of each factor on the landscape ecological risk. The results show that: 1) The spatial distribution of landscape ecological risks from 2002 to 2022 was characterized by "high concentration and low dispersion" from the perspective of temporal and spatial evolution, most of which were the low- and medium-low risk areas. The ecological risk level was basically high in the east and low in the west; The low- and medium-low risk areas increased by 491.89 km² over the past 20 years, while the high- and medium-high risk areas decreased by 926.17km². The ecological risk Basin was basically shifted from the high to low risk. The overall ecological risk shared the downward trend. 2) The landscape ecological risk was positively correlated in space, in terms of spatial and temporal distribution, with the significant spatial agglomeration and differentiation. The local agglomeration was mainly in the "H-H" and "L-L" pattern, with the gradually narrowing range of high-value agglomeration areas; The low ecological risk areas were distributed mainly in the middle and low terrain gradient on the vertical scale, while the high ecological risk areas were concentrated mostly in the range of low altitude and low slope gradient. The distribution index of high ecological risk increased with the decrease of altitude and the leveling of terrain, indicating the dominant distribution in space. 3) The landscape ecological risk was easily influenced by multi-dimensional factors. The explanatory power of natural environmental factors was generally stronger than that of human activities and climate conditions from the perspective of single factor. Each factor was relatively fluctuated in the different periods, but there was the strong influencing of topographic factors, human interference index, and annual average temperature; The interaction among factors showed a significant explanatory power enhancement at the level of factor detection, among which the interaction between natural environment and other factors was the strongest. The finding can also provide the useful reference for the management and control of ecological risks for the optimization of spatial layout.