Collaborative optimization methods for high-speed railway crew matching and rostering are investigated under multi-dimensional attributes. Comprehensive attribute systems are established for crew members—covering rank, route lines, skills, collaboration preferences, and available duty hours—and for crew duties, including duty grade, operating route, and duration. A two-stage optimization model coordinates crew matching and rostering, minimizing preference-order deviation and qualification disparity within teams and between teams and duties to balance qualification compliance and crew satisfaction. The complex multi-objective optimization model was transformed into a tractable single-objective linear programming model to enable efficient solution. The commercial solver GUROBI was employed to obtain exact solutions. Subsequently, sensitivity analyses were conducted on model parameters and problem scale to examine the relationships between parameter variations and optimization outcomes, as well as the impact of the first-stage optimization results on those of the second stage. Empirical validation with real-world depot data shows that, compared with manual planning, qualification matching and preference satisfaction improve by 47.7% and 3.7% in crew matching, and by 29.1% and 18.2% in scheduling, respectively. The approach delivers optimized solutions rapidly, offering an effective tool for intelligent, refined crew management.