In the last decades, effective cellulose degradation became a major point of interest due to the properties of cellulose as a renewable energy source and the widespread application of cellulases (the cellulose degrading enzymes) in many industrial processes. Effective bioconversion of lignocellulosic biomass into soluble sugars for ethanol production requires use of thermostable and highly active cellulases. The library of current cellulases includes enzymes that can work at acidic and neutral pH in a wide temperature range. However, only few cellulases are reported to be thermostable. In order to alleviate this, we have performed a hybrid approach for the thermostabilization of a key cellulase, Endoglucanase I (EGI) from Trichoderma reesei. We combined in silico and in vitro experiments to modulate the thermostability of EGI. Four different predictive algorithms were used to set up a library of mutations. Three thermostabilizer mutations (Q126F, K272F, Q274V) were selected and molecular dynamics simulations at room temperature and high temperatures were performed to analyze the effect of the mutations on enzyme structure and stability. The mutations were then introduced into the endoglucanase 1 gene, using site-directed mutagenesis, and the effect of the mutations on enzyme structure and stability were determined. MD simulations supported the fact that Q126F, K272F and Q274V mutations have a thermostabilizing effect on the protein structure. Experimental studies validated that all of the mutants exhibited higher thermostability compared with native EGI albeit with a decrease in specific activity.