PURPOSE: To setup procedures of quality assurance (QA) for implementing intensity modulated radiation therapy (IMRT) clinically, report QA procedures performed for one patient with prostate cancer. MATERIALS AND METHODS: P3IMRT (ADAC) and linear accelerator (Siemens) with multileaf collimator are used to implement IMRT. At first, the pos itional accuracy, reproducibility of MLC, and leaf transmission factor were evaluated. RTP commissioning was performed again to considers mall field effect. After RTP recommissioning, a test plan of a C-s haped PTV was made using 9 intensity modulated beams, and the calculated isocenter dose was compared with the measured one insolid water phantom. As a patient-specific IMRT QA, one patient with prostate cancer was planned us ing 6 beams of total 74 segmented fields. The same beams were used to recalculate dose in a solid water phantom. Dose of these beams were meas ured with a 0.015cc microionization chamber, a diode detector, films, and a narray detector and compared with calculated one. RESULTS: The pos itioning accuracy of MLC was about 1 mm, and the reproducibility was around 0.5 mm. For leaf transmission factor for 10 MV photon beams, interleaf leakage was measured 1.9% and midleaf leakage 0.9% relative to 10x10 cm2 open filed. Penumbra meas ured with film, diode detector, microionization chamber, and conventional 0.125 cc chambers howed that 80~20% penumbra width meas ured with a 0.125cc chamber was 2 mm larger than that of film, which means a 0.125 ccionization chamber was unacceptable for meas urings mall fields uch like 0.5 cm beamlet. After RTP recommissioning, the discrepancy between the meas ured and calculated dose profile for a small field of 1x1 cm2 size was less than 2%. The isocenter dose of the test plan of C-s haped PTV was meas ured two times with microionization chamber in solid phantom showed that the errors upto 12% for individual beam, but total dose delivered were agreed with the calculated within 2%. The transverse dose distribution meas ured with EC-L film was agreed with the calculated one ingeneral. The isocenter dose for the patient meas ured in solid phantom was agreed within 1.5%. Off-axis dose profiles of each individual beam at the position of the central leaf measured with film and array detector were found that at out-of-the-field region, the calculated dose underestimates about 2%, at inside-the-field the meas ured one was agreed within 3%, except some position. CONCLUSION: It is necessary more tight quality control of MLC for IMRT relative to conventional large field treatment and to develop QA procedures to check intensity pattern more efficiently. At the conclusion, we did setup an appropriate QA procedures for IMRT by a series of verifications including the measurement of absolute dose at the isocenter with a microionization chamber, film dosimetry for verifying intensity pattern, and another meas urement with an array detector for comparing off-axis dose profile.