Remanufacturing cost is one of the main basis for judging whether components have remanufacturing value. In order to improve the accuracy of remanufacturing cost prediction, a remanufacturing cost prediction model based on the smoothing index method was proposed.Through the model,the failure characteristics of waste parts were analyzed;combined with expert evaluation,the failure characteristics were quantified and normalized effectively. Finding Euclidean distance method through high-dimensional space was used to find the similarity.According to the similarity between the failure characteristic data and the historical completed remanufacturing data,the completed remanufacturing cost data with higher similarity were selected, and the smoothing index method was used to calculate the remanufacturing cost of scrap parts.So the prediction of parts remanufacturing costs was realized in the case of a small amount of historical sample data. Finally, an example verifies the accuracy and practicability of the model.