Real-world systems are made of interacting entities that evolve with time. Creating models that can forecast interactions by learning the dynamics of entities is an important problem in numerous fields. Earlier works used dynamic graph models to achieve this. However, real-world interactions are more complex than pairwise, as they involve more than two entities, and many of these higher-order interactions have directional components. Examples of these can be seen in communication networks such as email exchanges that involve a sender, and multiple recipients, citation networks, where authors draw upon the work of others, and so on. In this paper, we solve the problem of higher-order directed interaction forecasting by proposing a deep neural network-based model \textit{Directed HyperNode Temporal Point Process} for directed hyperedge event forecasting, as hyperedge provides native framework for modeling relationships among the variable number of nodes. Our proposed technique reduces the search space of possible candidate hyperedges by first forecasting the nodes at which events will be observed, based on which it generates candidate hyperedges. To demonstrate the efficiency of our model, we curated four datasets and conducted an extensive empirical study. We believe that this is the first work that solves the problem of forecasting higher-order directional interactions.