Objective: To determine which of the selected letters (F, A, S, M, R, P, E, and B) are more appropriate for a phonological verbal fluency task (PVFT) administered in Basque or Spanish to a group of Basque-Spanish bilinguals. Method: 57 healthy bilinguals adult from Basque Country, Spain. Inclusion criteria were 18–75 years old, completed primary school, a score of ≥27 on the Mini-Mental State Examination, a score of ≤4 on Patient Health Questionnaire–9, and fluent speaker of both Basque and Spanish. 61.4% of the sample were women, the mean age was 43.75 ± 16.99, and the mean years of education 15.77 ± 3.27. All participants completed the PVFT in both languages with order counterbalanced across participants. Results: Two-way ANOVAs for language and letter on number of words produced revealed significant differences between languages, F(1, 56) = 156.79, p< .0001, letters, F(7, 392) = 44.39, p< .001, and their interaction, F(7, 392) = 115.84, p< .0001. In Basque, participants produced most words to the letters B, A, and E, and in Spanish, participants produced most words to the letters P, M and R. Conclusion(s): These results suggest that cue letters need to be selected depending on the language of testing to facilitate optimal performance by participants. The present results indicate that the most appropriate letters in Basque are B, A, and E, and in Spanish, P, M and R. Selecting appropriate initial letters for PVFT is necessary to create comparable normative data for both languages.