Study region: The Delingha Basin is located on the northeastern Qinghai-Tibet Plateau, a climate-sensitive transitional zone. Study focus: Hydrochemical compositions of river water and groundwater contain abundant hydrological cycle information, while controlling mechanisms remain poorly quantified. This study integrates multivariate statistics including correlation analysis, Self-Organizing Maps (SOM), Principal Component Analysis (PCA) and Positive Matrix Factorization (PMF) with main ions and stable isotopes to decipher recharge source, hydrochemical characteristics and controlling factors for elucidating hydrochemical evolution mechanism. New hydrological insights for the region: River water and groundwater are recharged by mountain precipitation. SOM-PCA-PMF indicate hydrochemical characteristics vary spatially due to water-rock interaction. PMF further reveals distinct spatial patterns in five hydrochemical contribution factors (calcite and dolomite, gypsum, halite, mirabilite, cation exchange). A difference-similarity-difference pattern is observed along the flow path in the great contributions (>25 %) to river water and groundwater hydrochemistry. The hydrochemical compositions in the mid-rear alluvial-proluvial plain are governed by gypsum and halite for river water, but additional carbonate and cation exchange influences for phreatic water, indicating groundwater recharge via river seepage and lateral runoff. Further downgradient, similar contributing factors (calcite, dolomite, gypsum, cation exchange) indicate groundwater discharge. Consistent controls and stable isotopes confirm hydraulic connection between phreatic and confined water, resulting in the recurrence of different dominant contributions. These findings enhance hydrological cycle cognition and water resources allocation in arid/semi-arid areas