Language contact refers to a situation in which speakers of different languages meet and interact. There are three consequences of language contact. The first one is language shift, where one language is replaced by another. The second one is language maintenance, where languages coexist without any language change. The question of how language shift takes place and under what conditions coexistence of languages is possible has drawn wide attention of linguists as well as of theoretical researchers in other disciplines. Most of the previous mathematical studies of language dynamics have shown that there is majority advantage in the competition and that the population eventually becomes dominated by a single language that was spoken by a majority when the competition started. However, it is known that the contact of different languages could lead to the third consequence: the emergence of a new language, such as creole languages. Our study aims to model the spread of creole languages by using evolutionary game theory and determine the factors that drive the spread of creoles in the population. We consider an evolutionary game with individuals who speak either one of two ancestral languages or the creole that resulted from language contact between those two ancestral languages. Due to the formation process of creole, the creole speakers can communicate with speakers of other languages to some, but not full, extent. We assume that their frequencies change according to their payoffs that reflect how well they can communicate with others. Under this basic assumption, we varied (i) the population size and (ii) the difficulty of language change, to see their effects on the spread of the creole. Results show that the creole is more likely to spread if the population size is small enough for random drift to work well and/or if two ancestral languages are linguistically distant. We also find that finiteness of the population can either favor or disfavor the spread of creole depending on detailed conditions. Our results suggest that the contact between two distant languages and random drift drive the spread of creoles. It is known that many creole languages appeared in plantations where European owners brought labor forces from Africa as slaves. Our results match the homestead phase of these plantations, where European languages are distant from African ones, and where the population size was small.