The purpose of the study was to formulate a model to predict the performance of Iranian high school third-graders on the End of the Course Achievement (ECA) tests through their performance on the vocabulary tests, which were administered throughout the course. To meet this end, thirty two learners – aging seventeen to nineteen, all male – participated in the study which took nine months to complete. Their linguistic abilities were approximately at Intermediate-Mid level according to the ACTFL proficiency guidelines (1998). The sampling procedure was the intact group method. 333 lexical items were taught to the learners in the decontextualized paired-associate translation method. The classes were held two hours a week in a nine-month course of time. Six sets of vocabulary tests were administered and every learner’s average was calculated. The learners’ scores on the ECA tests and their average scores on the vocabulary tests were analyzed through the regression analysis procedure to derive a model that could reliably predict the learners’ ECA scores through their average performance on the vocabulary scores. The analysis yielded the following formula: (AVERAGE VOCABULARY × 0.713) + 2.871± [3.1].