Phenotypic variation analysis and comprehensive evaluation are important tools for selecting provenances of forest trees, which, in turn, is crucial for genetic improvement of forest trees. The study analyzed eight needle-leaf and fourteen tracheid trait indexes of ten provenances of Qinghai spruce forest stands. Multiple comparisons, correlation analysis, cluster analysis, and principal component analysis were used to evaluate the differences in needle-leaf and tracheid traits among the different provenances. The study found significant differences in the needle-leaf and tracheid traits among the provenances. All traits included 50 pairs of positive and 26 pairs of negative correlations. The coefficient of variation for the tracheid trait (18.86%) was higher than that for the needle-leaf trait (17.20%). A comprehensive evaluation of needle-leaf and tracheid traits was conducted using principal component analysis. The study demonstrated that the degree of variation in needle-leaf and tracheid traits among the provenances of Qinghai spruce was extensive, with a multitude of sources of variation. Therefore, it is crucial to strengthen the selection of good provenances in the early seed selection process. The results of the comprehensive evaluation can be utilized as a foundation for the selection of good provenances for carbon sequestration and timber properties.