Objective: To evaluate the association between self-rated evidence-based medicine (EBM) competencies and the prescription of drugs without scientific evidence against mild COVID-19 (present with any of the signs and symptoms of COVID-19 but who do not have shortness of breath, dyspnea, or abnormal chest imaging) among recently graduated physicians in Peru. Methods: We conducted an analytical cross-sectional study where we evaluated a non-probability sample of recently graduated physicians during June and July 2021 (end of second wave of COVID-19 in Peru). Self-rated EBM competencies were assessed by four domains (formulation of a clinical question, search, analysis, and application) using a Likert scale with scores from zero to four (“Very inadequate” = 0, to “Very Adequate” = 4), it was considered as “Adequate” if the score was three or four. In addition, the variable “General competence on EBM” was rated as “Adequate” if in all domains evaluated it presented an adequate self-rating. For the outcome, drug prescription, we considered the use of ivermectin, azithromycin, other antibiotics, hydroxychloroquine, dexamethasone, and anticoagulants (drugs with no efficacy demonstrated for patients with mild COVID-19). To assess the association, we used Poisson regression models with robust variances and obtaining crude (cPR) and adjusted (aPR) prevalence ratios with their 95% confidence intervals (95%CI). Results: Of a total of 239 physicians included 70.7% prescribed at least one drug without scientific evidence. A total of 51.1% reported adequate ratings in all evaluated domains of EBM. Self-rating the “Clinical Question Formulation” competency as adequate was associated with a lower frequency of prescribing medications for mild COVID-19 (aPR: 0.93; 95% CI: 0.91–0.95). While self-rating as adequate the competency of “Identify possible implications of investigations” was associated with an increase in the prescription of such drugs (aPR: 1.14; 95% CI: 1.09–1.20). Additionally, self-rating all domains as adequate were associated with less prescription (aPR: 0.93; 95% CI: 0.90–0.96). Conclusion: Seven out of ten recently graduated physicians prescribed some type of medication without scientific evidence to treat patients with mild COVID-19. Having adequate self-perceived EBM competencies was associated with a lower frequency of prescribing medications without scientific evidence to manage patients with mild COVID-19.