In the frame of the test of NISP instrument for ESA Euclid mission, the question was raised to perform a metrology measurement of different components during the thermal vacuum test of NISP instrument. NISP will be tested at Laboratoire d'Astrophysique de Marseille (LAM) in ERIOS chamber under vacuum and thermal conditions in order to qualify the instrument in its operating environment and to perform the final acceptance test before delivery to the payload. One of the main objectives of the test campaign will be the measurement of the focus position of NISP image plane with respect to the EUCLID object plane. To simulate the EUCLID object plane, a telescope simulator with a very well know focal distance will be installed in front of NISP into ERIOS chamber. We need to measure at cold and vacuum the position of reflectors installed on NISP instrument and the telescope simulator. From these measurements, we will provide at operational temperature the measurement of references frames set on the telescope simulator and NISP, the knowledge of the coordinates of the object point source provided by the telescope simulator and the measurement of the angle between the telescope simulator optical axis and NISP optical axis. In this context, we have developed a metrology method based on the use of a laser tracker to measure the position of the reflectors inside ERIOS. The laser tracker is installed outside the vacuum chamber and measure through a curved window reflectors put inside the chamber either at ambient pressure or vacuum pressure. Several tests campaigns have been done at LAM to demonstrate the measurement performance with this configuration. Using a well know reflectors configuration, we show that it is possible to correct the laser tracker measurement from the window disturbances and from the vacuum impact. A corrective term is applied to the data and allows retrieving the real coordinates of the reflectors with a bias lower than 30$\mu$m, which is lower than the laser tracker measurement uncertainties estimated at 60$\mu$m. No additional error term of the laser tracker measurement is observed when using the laser tracker with the curved window and in vacuum, comparing with a classical use of the laser tracker. With these test campaign, we have been able to demonstrate the possibility to use a laser tracker to measure in real time during a vacuum thermal test the position of different mechanical parts into a vacuum chamber with an accuracy better than 60$\mu$m.