This study reports the development of a scale to evaluate the speech act performance by intermediate-level Chinese learners of English. A qualitative analysis of the American raters' comments was conducted on learner scripts in response to a total of 16 apology and request written discourse completion task (WDCT) situations. The results showed that in content, these raters attended to features, including "amount of information," "politeness," "clarity," and "relevance." For form, they focused on "grammar," "phrasing," and "word choice." These features were then incorporated into the scale design, resulting in a scoring rubric consisting of two 10-point subscales, each of which contains five levels with descriptors for each performance level. For scale validation, seven experienced American writing teachers in Taiwan evaluated 84 e-mail requests written by the Chinese learners to their professors. Reliability and validity of the rating scale were explored through the FACETS analysis. The results showed that the scale can be used in a reliable and valid way. The study concludes with a discussion of the usefulness of this rating scale and proposes directions for future research.