This paper investigates the impact of the tonal factor on the realization of diphthongs in Standard Mandarin. We employed a set of Generalized Additive Mixed Models (GAMMs) to respectively test whether and how tones and fundamental frequency (f 0) influence the realization of the diphthong /ai/ in a journalistic speech corpus of Standard Mandarin. Results show, in general, the realizations of /ai/ to significantly differ with respect to tones, due to different heights and contours of f 0. A high tone results in a more closed and more front /ai/ and a low tone results in a more open and more back realization; a rising tone strengthens the dynamic features of /ai/ while a falling tone weakens such features and leads to monophthongization. The results suggest that the relation between f 0 and vowel realization, found in monophthongs, is equally applicable to diphthongs: f 0 is negatively correlated with F1 and positively with F2.