Abstract The primary objective of this work was to investigate the results and the possible mechanisms of oil recovery in a huff-n-puff style improved oil recovery (IOR) field pilot using nanoparticle assisted CO2 injection. A secondary objective was to study the sensitivity of the process to injection volume of nanoparticles and gas, the type of injected gas, soaking period, and the timing of IOR to maximize net present value. An Eagle Ford shale well was produced for 526 days before 167-barrels of nanoparticle treatment and 160-tons of CO2 were injected in 11 cycles into the well, shut-in for 5 days and then put back on production. A simulation study was conducted using a fully coupled geomechanical compositional fracturing and reservoir simulator using data from the pilot well. The primary production was history matched for the fractured horizontal well and the huff-n-puff process with nanoparticle and CO2 injection was simulated followed by a shut-in period. The simulated production after shut-in and the incremental oil recovery was compared with field measured data. The pilot test results clearly show that there is a significant oil rate increase after the nanoparticle and CO2 are injected. Lab results show that nanoparticles can lower the interfacial tension between the water and oil and alter the rock wettability to a preferential water-wet state, which is beneficial for oil production. The simulation studies show that CO2 injection alone results in smaller improved oil recovery and predicts a smaller oil recovery than in the field. This suggests that both the nanoparticles and gas play an important role in increasing the relative permeability to oil and improving oil recovery. Results from the sensitivity study show that larger injection volumes of nanoparticles and gas result in higher oil recovery. Among different injection gases simulated, in this oily window of the Eagle Ford shale, ethane gives the highest oil recovery followed by CO2, methane, and nitrogen. A longer soaking period after the injection also helps to increase oil recovery. It is also shown that it may be better to perform IOR at an earlier stage of primary production to maximize the cumulative oil recovery. Our field and simulation results provide operators with significant new insights into the design of an IOR process that uses nanoparticles with CO2 injection. The integration of field pilot test data with realistic compositional geomechanical reservoir simulation for the first time provides a quantitative estimate of the improvement in oil recovery and insights into the possible mechanisms of oil recovery.