Systemic Epstein-Barr virus-positive T-cell and natural killer (NK)-cell lymphoproliferative diseases of childhood are a group of lethal diseases mostly affecting children and young adults. The Ohshima Grading System and the 2017 World Health Organization (WHO) classification have been used for classifying this spectrum, but these systems have not been validated externally and compared. Therefore, we examined 36 cases of systemic Epstein-Barr virus-positive T-cell and NK-cell lymphoproliferative diseases of childhood with long-term follow-up, from Southwest China, to systematically summarize the clinicopathologic features and to validate and compare the Ohshima Grading System and the 2017 WHO classification in discrimination ability, predictive accuracy, concordance indices, and explained variation. Clinically, our cohort showed severe manifestations and poor prognoses. Morphologically, the hematopoietic and lymphoid specimens showed proliferation of small-sized to medium-sized bland-looking lymphocytes that might mask disease severity, whereas other extranodal lesions showed a disorganized to obliterated architecture infiltrated by medium-sized to large-sized, subtle to obvious atypical cells, which may mimic extranodal NK/T-cell lymphoma. Immunophenotypically, our cases mainly originate from CD8 αβ T cells. Therefore, clinical and pathologic features should be equally considered to avoid missed diagnosis or misdiagnosis. In addition, the 2017 WHO classification shows a flexible grasp of pathologic features, thus classifying some cases (polymorphic and monoclonal cases with fulminant course) more reasonably; thereby, it showed statistically improved results compared with the Ohshima Grading System. However, underestimating the risk of some polyclonal cases and imprecisely discriminating monoclonal cases at diagnosis are common dilemmas in both systems. Therefore, the construction of a comprehensive grading algorithm for improved prognostic value and precise diagnosis requires additional studies.