Whole blood transfusions are an essential treatment modality during rehabilitation of stranded sea turtles, however, standardized protocols for transfusions are not available in reptile medicine. The objective of this study was to optimize a cross-matching protocol for sea turtle blood transfusions. Fresh venous blood samples from 15 turtles (n = 14 green turtles, Chelonia mydas [Cm]; n = 1 loggerhead sea turtle, Caretta caretta [Cc]) were tested using a temperature-appropriate (i.e., reflecting body temperature), time-sensitive protocol in 26 reactions using two procedures for cross-match evaluation at 30 and 60 min at ambient air and water bath temperature. There were no significant differences between both protocols at 30- and 60-min incubation times or between microscopic evaluations at 2 or 5 min. The major cross-match identified 7/22 incompatible Cm–Cm reactions as observed by microscopic erythrocyte agglutination. Minor cross-matches resulted in 6/22 incompatible Cm–Cm reactions. About half of all Cm–Cm reactions (12/22) were compatible by major and minor cross-match. All Cc–Cm reactions (4/4) were incompatible. A higher than expected proportion of incompatible Cm cross-matching reactions suggests preexisting antibodies to nonself erythrocyte antigens in this species, or other factors promoting erythrocyte aggregation or agglutination. Preliminary data across Cm and Cc suggests cross-species incompatibility. These results indicate that sea turtles may react to donor erythrocytes even at the first transfusion. Concurrent major and minor cross-matching using the proposed protocol at 30-min incubation at room temperature should be considered a necessary and effective way to test for patient and donor incompatibilities in sea turtles.