Understanding the dynamics of the soil-vegetation-atmosphere (SVA) system, particularly in arid and semiarid regions, remains challenging due to its intricate and interdependent nature. This system creates problems for rangeland administration, such as insurance and risk management. This paper focuses on the complex interactions within the SVA system, particularly on rangeland ecosystems in Spain's semiarid and arid regions. By employing multifractal detrended cross-correlation analysis (MFCCA), we explore the joint behavior of key variables, including precipitation (PCP), evapotranspiration (ETP), aridity index (Arid. I.), soil water availability (SWA), biomass (Bio), and normalized difference vegetation index (NDVI). Analyzing a 20-year data series from Madrid and Almeria provinces, we reveal distinct patterns in the studied variables' persistence, multifractality, and asymmetry. Notably, the differences in the generalized Hurst exponents (hxy${{{\mathrm{h}}}_{xy}}$(q)) between Madrid and Almeria for SWA with NDVI, SWA with Bio, and NDVI with Bio underscore distinct interactions in these regions. Moreover, multifractal analyses unveil differences in the complexity of joint variables' behaviors in the two regions. Almeria exhibits higher multifractality across variables, indicating more complex and variable environmental interactions, likely due to its more arid conditions. These findings suggest that Almeria has more sensitivity to changes, requiring adaptive management strategies, while in Madrid, water availability and related variables play a more dominant role in driving vegetation dynamics. These findings shed light through MFCCA on the nuanced dynamics of rangeland ecosystems in semiarid and arid regions, emphasizing the importance of considering complexity-based approaches to understand the intricate interplay among key variables in the SVA system. We show multifractal detrended cross-correlation analysis as a tool to describe joint and complex dynamics. Differences are found in the ecological interaction between arid and semiarid rangelands. Cross-correlated and average univariate analyses provide differences in multifractal parameters. Almeria shows higher multifractality, indicating more complex and variable environmental interactions. Soil water availability patterns influence normalized difference vegetation index and biomass dynamics more in Madrid than in Almeria. Ecosystems are highly interacting systems, where some changes can produce surprising outcomes (complex systems). This study explores the relationships among various ecological components of rangelands in Madrid and Almeria (Spain) to aid rangeland managers. Using multifractal analysis, which studies the behavior of time series, we compare the behavior and the influence of different environmental factors like precipitation or biomass. Our findings show that Almeria, with its more arid conditions, has more complex and variable environmental interactions and is more sensitive to small changes. While in Madrid, soil water availability has a stronger influence on vegetation growth. These insights suggest that rangeland management strategies should be tailored to the specific conditions of each region to improve sustainability and productivity. Almeria should prioritize enhancing soil water retention and in Madrid, managers should focus on monitoring larger scale climatic events.