Grinding is an abrasive process mostly used in finishing operations to provide low roughness and narrow limits of form and dimensioning to the workpiece. Due to the large amount of heat generated by friction between the abrasive and the workpiece in this process, the use of large volumes of coolant is encouraged to avoid thermal damage, such as burning and hardness variation caused by subsurface damage. On the other hand, environmental impacts and human health problems caused by coolants have been a key issue toward sustainable manufacturing, mainly because of the chemistry behind them. Thus, is important to seek for strategies to reduce the volume of fluids and their risks as well as guarantee grinding efficiency. One machining strategy is the minimum quantity of lubricant (MQL) technique, which is well consolidated over the past 25 years and one that uses low volumes of fluid mixed with compressed air flow, as well as provides less waste. However, it has generally been reported that sludge formed during grinding is forced into the wheel pores, consequently clogging its pores, thereby reducing the wheel cutting potential and its performance. A possible solution for this problem is to use an auxiliary compressed air system to clean the grinding wheel surface during machining, since the MQL conventional system is not able to clean it. In this context, this work evaluated the performance of the MQL technique with an auxiliary cleaning of the grinding wheel cutting surface in relation to the conventional cooling techniques (flood cooling) during a cylindrical plunge grinding of N2711 steel. N2711 steel is widely employed in manufacturing of molds for plastic injection processes and is one of steels more susceptible to grinding burn. The following output parameters were used to assess the performance: surface roughness, roundness, microhardness, grinding power, and grinding wheel wear. The results showed that the MQL technique, in addition to the environmental and economic advantages achieved, provided superior workpiece quality, and lower power consumed compared to the flood technique. The MQL technique proved to be an alternative method compared to the conventional technique under the conditions investigated. Also, the Malkin’s model was used to predict the grinding ratio (G-ratio) based on the experimental data obtained in this work. After regression analysis, the model predicted the G-ratio from the specific material removal rate and the cutting speed with a satisfactory accuracy of approximately 92%.