We propose a novel nonparametric approach for estimating the production frontier based on a data-fitting technique. The proposed approach allows for stochastic noise and provides decision-makers with a general and flexible estimation procedure to support and facilitate the decision-making process in the stochastic context. A major feature of the proposed approach is that the estimation procedure is completely nonparametric and easy to implement. Similar to other existing nonparametric approaches, our proposed approach results in an estimate of the piece-wise linear production frontier. In contrast to the existing ones, the evaluation of each data point is performed within a unit-specific data range. We also propose a naive method for determining the data ranges of each data point. The performance of our proposed approach is examined using various simulated scenarios. For each scenario, we compare our proposed approach with the existing methods, including the data envelopment analysis (DEA), the stochastic nonparametric envelopment of data (StoNED), and the stochastic frontier analysis (SFA). The simulation results suggest that our approach performs better than the existing methods in the single input and single output case. Our proposed approach can also be easily extended to a multi-input setting. Moreover, the proposed naive method on data ranges also shows its flexibility and usefulness in the simulated examples.