In this research paper, we present a novel language learning/training and assessment system that helps people to learn and practice a new language independently at low cost. To achieve an independent-learning workflow, we explore the use of real-time speech recognition, language translation, speech synthesis, and language intelligibility assessment technologies to provide automatic assessment and instant feedback of language-speaking performance. We also propose and adopt an objective assessment methodology that determines the intelligibility based on outcome of speech recognition. Our experimental results demonstrate that the proposed system can sufficiently analyze the intelligibility of one's speaking, accurately identify mispronounced words, and define a feedback that localizes and highlights errors for helping continuous practice toward perfection. [For the complete proceedings, see ED590271.]