To understand an indirect request (e. g.“I want to know the time.” to ask for the time), the context is very important. The purpose of this study was to investigate how important contextual information was, and to see how it affected understanding indirect requests. First, in order to identify the important contextual information, we asked subjects through a questionaire to make indirect requests in various circumstances. what being reffered to frequently in the indirect requests (e. g. speaker's goal: “I want to know the time.”) suggested the premise for felicitous requests. We considered the information about such premise (e. g. A speaker wants to know the time.) as being the important contextual information. From the results of the questionaire: speaker (S)'s goal, S's condition (unable to achieve the goal), S's expectancy for hearer (H)'s help, H's condition (able to achieve the goal for S), H's attitude (cooperative) were considered to be the premise for (S) requesting H to do some action. Then, the effects of contextual information (information about S's goal, S's condition, H's condition, H's attitude) on understanding the illocutionary force (Exp. 1) and the content (Exp. 2) of indirect requests were investigated. In Exp. 1, a contextual information, an indirect representation of a request, then a direct representation of the request were presented on CRT consequently. The subjects' task was to decide if the illocutionary force of the indirect representation was the same as that of the direct one, considering the contextual information. The response and response time were recorded. In Exp. 2, several contextual informations for a request were consequently presented on CRT. The subjects' task was to suspect the content of the request, taking the information into account one by one. The subjects pressed the key and answered as soon as they confirmed their suspect. The response and response time were recorded. The results showed that to understand the illocutionary force, the information about S's goal and H's attitude were effective; also to understand the content of requests, the information about S's goal had little effect, the information about S's condition and H's condition had a greater effect, while information about H's attitude had the greatest effect.