The success of recommender systems has made them the focus of a massive research effort in both industry and academia. The aim of most recommender systems is to identify a ranked list of items that are likely to be of interest to users. However, there are several applications such as trip planning, where users are interested in package recommendations as collections of items. In this paper, we consider the problem of recommending the top-k packages to the active user, where each package is constituted with a set of points of interest (POIs) and is under user-specified constraints (time, price, etc.). We formally define the problem of top-k composite recommendations and present our approach which is inspired from composite retrieval. We introduce a scoring function and propose a ranking algorithm for solving the top-k packages problem, taking into account the preferences of the user, the diversity of recommended packages as well as the popularity of POIs. An experimental evaluation of our proposal, using data crawled from Tripadvisor demonstrates its quality and its ability to provide relevant and diverse package recommendations