A new Thermal Ionization Mass Spectrometer (TIMS) analytical procedure was developed to measure with high accuracy the zirconium isotope abundance and concentration without molybdenum correction in nuclear samples. A zirconium selective separation using UTEVA column was used before the TIMS measurement in order to remove all possible isobaric interferences (Mo, Y, Nb and Ru) in the solution. The separated solution was then deposited onto a filament previously outgassed for 4 hours to reduce the molybdenum traces in the filament. Then, the filament was introduced into the TIMS source and was maintained for 4 hours at a current of 5 A (ionization filament) and 1.2 A (evaporation filament) in order to eliminate all molybdenum trace present in the filament and in the sample. This methodology allowed the zirconium isotope abundance determination without using the molybdenum interference correction equation. The analytical results obtained with a natural solution after separation is in good accordance with the reference values: bias lower than 0.16 % for the 90 Zr, 91 Zr, 92 Zr and 94 Zr isotope abundance, and a bias of ~0.8 % for the minor isotope (96 Zr) were observed. This zirconium abundance determination methodology was used on three nuclear samples. Then, the calibration of a 91 Zr spike solution was performed by reverse isotope dilution using a SPEX standard zirconium solution. The zirconium concentration determination was then performed on the three nuclear samples by isotope dilution using the SPEX standard zirconium solution and the 91 Zr spike solution. The final uncertainties were estimated to be lower than 1 %.