Coreference resolution is an NLP task to find out whether the set of referring expressions belong to the same concept in discourse. A multi-pass sieve is a deterministic coreference model that implements several layers of sieves, where each sieve takes a pair of correlated mentions from a collection of non-coherent mentions. The multi-pass sieve is based on the principle of high precision, followed by increased recall in each sieve. In this work, we examine the portability of the multi-pass sieve coreference resolution model to the Indonesian language. We conduct the experiment on 201 Wikipedia documents and the multi-pass sieve system yields 72.74% of MUC F-measure and 52.18% of BCUBED F-measure.