This study compares two instructional methodologies: extensive reading plus output activity and intensive reading plus grammar-translation with regard to improvement of non-English major EFL learners' reading comprehension and fluency development in a Japanese junior college. It identifies the minimum number of words an individual should read for improvement, developing a classroom instructional model that can be implemented as a national policy. The results indicated that 50,000 words should be read to be as effective as conventional instructional methodology. Reading fluency can be improved regardless of instruction types or words read. The study time could peak before tertiary study, aligning with extrinsic motivation to study English. The results of questionnaires enquiring about their attitudes towards English learning and extensive reading are also discussed.