The NanoMQL technique is used to overcome the limitations of wet grinding due to economic and ecological problems. The performance measure is largely influenced by the process parameters such as table speed, depth of cut, air pressure, coolant flow rate and nanofluid concentration. In this paper, the performance of NanoMQL technique in terms of surface roughness was evaluated for hard and soft EN31 steel. The Experiments were conducted by response surface methodology (RSM) using statistical software to develop regression model of surface roughness and optimization was carried out using Jaya algorithm. The result shows that lowest value of surface roughness was obtained for NanoMQL of hard steel in comparison with soft steel under grinding environ-ments such as wet, MQL and NanoMQL. Hence to improve the performance of soft steel, the modeling and optimization of surface roughness were carried out. The significant parameters were considered for model development and validity of model determined through ANOVA (Analysis of variance). Lastly, the optimal values were determined using Jaya algorithm for minimum surface roughness and the percentage error observed to be close with the experimental test.