Decay heat measurements, called the MERCI experiment, were conducted at Commissariat à l'Energie Atomique (CEA)/Saclay to characterize accurately residual power at short cooling time and verify its prediction by decay code and nuclear data. The MOSAÏC calorimeter, developed and patented by CEA/Grenoble (DTN/SE2T), enables measurement of the decay heat released by a pressurized water reactor (PWR) fuel rod sample between 200 and 4 W within a precision of 1%. The MERCI experiment included three phases. At first, a UO2 fuel rod sample was irradiated in the CEA/Saclay experimental reactor OSIRIS. The burnup achieved at the end of irradiation was [approximately]3.5 GWd/tonne. The second phase was the transfer of the fuel rod sample from its irradiation location to a hot cell, to be inserted inside the MOSAÏC calorimeter. It took 26 min to carry out the transfer. Finally, decay heat released by the PWR sample was measured from 27 min to 42 days after shutdown. Postirradiation examinations were performed to measure concentrations of some heavy nuclei (U, Pu) and fission products (Cs, Nd). The decay heat was predicted using a calculation scheme based on the PEPIN2 depletion code, the TRIPOLI-4 Monte Carlo code, and the JEFF3.1.1 nuclear data file. The MERCI experiment analysis shows that the discrepancy between the calculated and the experimental decay heat values is included between -10% at 27 min and +6% at 12 h, 30 min after shutdown. From 4 up to 42 days of cooling time, the difference between calculation and measurement is about ±1%, i.e., experimental uncertainty. The MERCI experiment represents a significant contribution for code validation; the time range above 105 s has not been validated previously.