The purpose of this project is to adapt Named Entity Recognition (NER) to the conflict research domain. The conflict domain is represented by data from the Uppsala Conflict Data Program (UCDP). We want to build up the starting steps for automated information extraction from the source data in this domain. As there are no existing NER data sets for this domain, at least testing data must be created to enable testing of NER adaption in the domain. We first survey available NER data sets. From these, we chose to adapt ACE-2 Version 1.0 data and its guidelines to conform to our wanted NER entities for the conflict domain. This adapted data we then use for training a NER model, making this project a cross-domain adaption task, as we evaluate on testing data from a different domain. Using the adapted guidelines we annotate a data set of 150 documents from the UCDP data, which we use for evaluating performance in the conflict domain. We experiment with two strategies for NER for this cross-domain adaption task. First, using gazetteers and various configurations. Second, using enrichment of training data with conflict data documents, and we also map the learning curve for adding conflict domain related documents to the ACE-2 Version 1.0 training data. This project serves as an example of adapting to a new domain for NER using pre-existing training data, and show adaption approaches which can be further built upon for information extraction in this domain.