Populus deltoides is one of the primary tree species for bioenergy production in temperate regions. In arid/semi-arid northern China, the scarcity of water and nitrogen significantly limits the productivity of poplar plantations. The identification of relevant molecular markers can promote the breeding of resource-efficient varieties. In this study, 188 genotypes of P. deltoides from six provenances served as experimental material. Genetic differentiation analysis, analysis of molecular variance (AMOVA), principal coordinate analysis (PCoA), unweighted pair group method with arithmetic mean (UPGMA) clustering, and genetic structure analysis were performed using selected simple sequence repeat (SSR) markers. Based on these analyses, the association analysis of water-use efficiency (WUE) and nitrogen-use efficiency (NUE) were conducted using general linear model (GLM) and mixed linear model (MLM) approaches. The results showed that 15 pairs of SSR primers successfully amplified across all 188 individuals, with an average of 7.33 alleles (Na) observed per primer pair. The polymorphism information content (PIC) ranged from 0.060 to 0.897, with an average of 0.544, indicating high genetic diversity in the selected markers. The average inbreeding coefficient intra-population (Fis), inbreeding coefficient inter-population (Fit), and inter-population genetic fraction coefficient (Fst) values were 0.005, 0.135, and 0.132, respectively, indicating high heterozygosity, substantial inbreeding within populations, and moderate genetic differentiation, with an average gene flow (Nm) of 1.964, suggesting substantial gene flow between populations. Additionally, molecular variance was primarily within individuals (84.12%). Genetic structure analysis revealed four subgroups, with some degree of genetic admixture among the provenances. In the GLM model, 11 markers were significantly associated with five traits (p < 0.05), with an average contribution rate of 15.82%. Notably, SSR132 and SSR143 were significantly associated with multiple traits (p < 0.05). The MLM model identified two markers (SSR47 and SSR85) significantly associated with ground diameter (p < 0.05) and one marker (SSR80) significantly associated with NUE (p < 0.05). This study identifies loci associated with WUE and NUE, laying a foundation for future genetic improvement and marker-assisted breeding strategies in poplar.