Assay and analysis procedures were developed for nondestructive fissile isotopic measurement of mixed 233U-235U fuel samples. For 233U, the number of delayed neutrons released per fission is about half that for 235U, although the number of prompt neutrons is approximately the same. By separately counting prompt and delayed neutrons released by a sample exposed to neutron irradiation, the amounts of 233U and 235U present in the sample can be determined. Equations of delayed and prompt neutron counts versus 233U and 235U contents are solved simultaneously for the 233U and 235U contents of a sample. Eleven samples containing mixtures of 233U and 235U from no 233U to nearly 100% were prepared and assayed in prompt and delayed neutron assay devices. Constants for calibration equations were fitted to data from nine of the samples. The maximum differences between counts calculated by the calibration equations and measured counts were 2.3% for delayed neutrons and 1.2% for prompt neutrons, indicating a good selection of the form for the calibration equations. The two remaining samples were treated as unknown, and the uranium contents of these samples were determined by simultaneously solving the two calibration equations. The maximum difference between measured 233U or 235U content and actual content for either sample was 1.5%.