Abstract : In this paper we address the problem of predicting SPARQL query performance. We use machine learning techniques to learn SPARQL query performance from previously executed queries. Traditional approaches for estimating SPARQL query cost are based on statistics about the underlying data. However, in many use-cases involving querying Linked Data, statistics about the underlying data are often missing. Our approach does not require any statistics about the underlying RDF data, which makes it ideal for the Linked Data scenario. We show how to model SPARQL queries as feature vectors, and use k-nearest neighbors regression and Support Vector Machine with the nu-SVR kernel to accurately predict SPARQL query execution time.