CO2 injection in depleted reservoirs has been largely employed over the past years as an effective process for oil and gas enhanced recovery. More recently, the injection of CO2 in geological deposits is considered a viable alternative to reduce greenhouse gases in the atmosphere. In these two scenarios, the injected CO2 interacts with the rock deposit altering some petrophysical and geomechanical properties. One of the main deposits is carbonate reservoirs. The objective of this research is to improve the knowledge regarding the mechanical effects of CO2 injection on carbonate rocks. Changes to the rock pore structure may change oil flow, cause reservoir compaction and containment issues. In order to improve the insight on the interaction between CO2 and carbonate rocks, laboratory tests and numerical models were carried out in this study. For the laboratory tests, a solution of liquid CO2 and deionized water was injected through Indiana Limestone and coquina samples. In addition, mechanical tests were executed before and after CO2 injection, to evaluate the effects on the mechanical properties from the carbonate rocks. The mechanical characterization was conducted by performing uniaxial and triaxial compressive tests. From the laboratory results, significant reduction on the unconfined compressive strength and on Young s moduli from the coquina and Indiana Limestone samples when subjected to the CO2 injection test was observed. In addition, an increase in porosity was also noted from MicroCT scans and porosimetry measurements on the coquina sample after the injection of CO2 solution. These results were used for the numerical model calibration and validation. Regarding the numerical model, the Discrete Element Method (DEM) was adopted. To accurately simulate the material behavior, a methodology encompassing the contact model, the sample generation and the calibration procedure is presented. The calibration of the DEM parameters used the results from the mechanical tests on the rocks prior and after the dissolution tests. The numerical model was able to accurately simulate the uniaxial and triaxial tests on the Indiana Limestone and coquina samples. In addition, the methodology presented for the DEM model managed to accurately reproduce the mechanical degradation due to CO2 injection on Indiana Limestone and to a lesser degree, on coquina samples.