Background/Objectives: Genetic variation in human leukocyte antigen (HLA) genes may contribute to inter-individual differences in infectious-disease susceptibility and clinical outcomes. This study aimed to determine the genotype frequency of HLA Class I and Class II loci in patients with COVID-19 in Saudi Arabia and to examine their associations with survival and clinical severity. Methods: A prospective observational study was conducted at King Faisal Specialist Hospital and Research Centre (KFSH&RC), Riyadh, Saudi Arabia, from January 2022 to December 2023. Genomic DNA was extracted, and polymerase chain reaction-sequence specific oligonucleotide (PCR-SSO) testing was performed to screen HLA genetic variation. Patients were grouped by survival status (recovered or deceased) and clinical severity: Stage A (asymptomatic), Stage B (mild), Stage C (moderate), and Stage D (severe). Results: In total, 123 patients with COVID-19 were included; 102 (82.9%) recovered and 21 (17.1%) died. ICU admission was more frequent among deceased patients than among recovered patients (95.2% versus 51.0%, p = 0.0001). At the locus level, HLA-DPB1 represented the largest proportion of HLA calls (21%). In call-position-specific allele-group analyses, B*15 in Allele 1 (14.3% versus 1.0%, p = 0.016), C*06 in Allele 2 (42.9% versus 18.6%, p = 0.023), DRB1*10 in Allele 1 (19.0% versus 4.9%, p = 0.045), and DQB1*05 in Allele 1 (33.3% versus 11.8%, p = 0.021) were significantly more frequent among deceased patients, whereas DQB1*03 in Allele 1 was significantly more frequent among recovered patients (45.1% versus 14.3%, p = 0.013). Severity analyses showed call-position-specific differences involving C*15, C*06, B*14, B*39, B*53, and DQB1*03. Vaccination status did not differ significantly by survival status or across the four clinical severity stages. Conclusions: Selected HLA Class I and Class II allele groups may be associated with COVID-19 survival and clinical severity patterns in this Saudi cohort. These findings should be interpreted cautiously given the cohort size and call-position-specific nature of the analyses.