In this paper, the location routing problem (LRP) with outsourced delivery is studied. The objective of LRP is to determine the best location of depots and vehicle routes that minimizes the total costs. In classical LRP, all products are delivered with non-outsourced delivery, which is the delivery mode that all customers are included in delivery routes. The non-outsourced delivery is efficient if customers are densely located. However, if customers are sparsely located, the outsourced delivery is more efficient. In outsourced delivery, products are delivered to customers directly by third-party delivery services. In the study the outsourced delivery mode to the LRP model so that a suitable transportation mode is selected. Introducing this model makes the LRP more complicated. To solve the problem efficiently, a clustering-based heuristic is proposed. Test instances based on LRP benchmark instances were solved by the algorithm. Results of computational experiments showed the proposed algorithm can found 20% better solution than MIP within an hour for problems with a hundred customers.