This study investigated prosodic aspects of native and non-native (native Japanese, NJ) speech in American English. Since Japanese is characterized as mora-timed and English as stress-timed, it was hypothesized that durational contrasts among syllables in English spoken by NJ speakers may not be as large as those in native speech. In addition, it was hypothesized that a wider fundamental frequency (F0) range may be found in utterances produced by NJ speakers than in native speech, because NJ speakers may rely on F0 to indicate stress due to pitch accent in Japanese. The participants were NJ adults and children (16 each), and age-matched native English (NE) adults and children (16 each). They repeated ten phrases/sentences after a model. Durations of each syllable and the whole utterance were measured, and the F0 range in each syllable and the whole utterance was calculated in semitones for three out of the ten phrases/sentences. The results indicated that absolute durations of syllables and utterances tended to be longer in the NJ speakers’ utterances than in the NE speakers’. Children’s utterances (NE and NJ combined) also tended to be longer than adults’. The F0 range was larger in the NJ speakers’ utterances than in the NE speakers’ in one target sentence. In addition, it was found that the duration of function words was proportionately longer in the NJ speakers’ utterances than in the NE speakers’ and that the F0 range was greater in the NJ adult speakers’ utterances than the NE adult speakers’ in content words. The results of this study partially support the hypothesis that NJ speakers rely on pitch difference to indicate stress in English. It is also possible that function words were not as reduced in nonnative speech as in native speech. Differences were found between adults and children, indicating that there may be developmental changes in prosodic aspects.