The immune mechanisms of long coronavirus disease 2019 (COVID) are not yet fully understood. We aimed to investigate the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)-specific memory immune responses in discharged COVID-19 patients with and without long COVID symptoms. In this cross-sectional study, we included 1041 hospitalized COVID-19 patients with the original virus strain in Wuhan (China) 12 months after initial infection. We simultaneously conducted a questionnaire survey and collected peripheral blood samples from the participants. Based on the presence or absence of long COVID symptoms during the follow-up period, we divided the patients into 2 groups: a long COVID group comprising 480 individuals and a convalescent group comprising 561 individuals. Both groups underwent virus-specific immunological analyses, including enzyme-linked immunosorbent assay, interferon-γ-enzyme-linked immune absorbent spot, and intracellular cytokine staining. At 12 months after infection, 98.5% (1026/1041) of the patients were found to be seropositive and 93.3% (70/75) had detectable SARS-CoV-2-specific memory T cells. The long COVID group had significantly higher levels of receptor binding domain (RBD)-immunoglobulin G (IgG) levels, presented as OD450 values, than the convalescent controls (0.40 ± 0.22 vs 0.37 ± 0.20; P = .022). The magnitude of SARS-CoV-2-specific T-cell responses did not differ significantly between groups, nor did the secretion function of the memory T cells. We did not observe a significant correlation between SARS-CoV-2-IgG and magnitude of memory T cells. This study revealed that long COVID patients had significantly higher levels of RBD-IgG antibodies when compared with convalescent controls. Nevertheless, we did not observe coordinated SARS-CoV-2-specific cellular immunity. As there may be multiple potential causes of long COVID, it is imperative to avoid adopting a "one-size-fits-all" approach to future treatment modalities.