In JT-60U, D-D experiments have been performed since July 1991. Although tritium, which is produced by the D-D reaction, is programmed to be exhausted through pumping systems, a part of the tritium is trapped in the graphite inner wall. To evaluate the tritium retention in the graphite wall of JT-60U, several sample tiles were removed from JT-60U in November 1993. The retained tritium distribution in the graphite tiles was measured by means of burning sample pieces, converting the tritium into HTO and scintillation counting the resultant water. The result of this measurement showed that tritium of 1.6 × 10 10 Bq still remained in the graphite first wall. This value corresponds to approximately 50% of the tritium generated from July 1991 to October 1993 in JT-60U. The tritium remaining in the divertor tiles was in higher concentration than that in the first wall. In particular, the tritium concentration peaked in the tiles between the strike points. Tritium at the strike points, where an electron or ion strikes directly, was in low concentration.