In real-world, managers and decision makers (DMs) of unconventional water resources (UWR) such as stormwater, reclaimed and flood water resources often face various uncertainties and work in changing environment. These uncertainties and changing factors derive from imprecise and uncertain information as well as subjective decision uncertainty associated with UWR management. Consequently, in multi-criteria decision making (MCDM) and assessment for UWR management, there exist uncertainties on criteria weights (CWs) and performance values (PVs). Additionally, qualitative criteria in multi-criteria system are often difficult to quantify precisely, making conventional MCDM models difficult to handle. To this end, considering uncertain PVs and CWs, integrated stochastic MCDM framework is developed for solving UWR management where quantitative and qualitative criteria are mixed. The framework contains: 1) novel stochastic multi-criteria acceptability analysis (SMAA) model based on modified grey relational analysis (GRA) and bidirectional projection (BP); 2) SMAA-Ordinal (SMAA-O) model for handling qualitative criteria; 3) coordinated CWs to compromise conflict of multiple criteria and simulate potential uncertainty; 4) decision-making (DM) risk quantification by risk analysis index. The framework is demonstrated by two cases of reclaimed water utilization (RWU) assessment and flood water control (FWC). Results show that framework can effectively handle uncertain information and qualitative criteria quantification. Parameter significance is analyzed to disclose its impact on DM uncertainty. Effectiveness analysis is conducted to demonstrate superiority of novel SMAA-GRBP coupled SMAA-O. Results indicate that framework provides feasible way for understanding stochastic essence of DM process, uncertainty propagation, probabilistic ranking information and releasing robust decisions with quantified decision risk at hand.