Self-mixing interferometry (SMI) is a contact-less sensing technique which is used for different metrological and biomedical applications. SMI phenomena occurs when laser beam is reflected back into the laser cavity by a remote target surface and mixes with the originally generated optical beam. The resultant SMI signal carries the information of the target (for e.g., vibration, displacement etc.) which can be retrieved by processing it. Conventionally, self-mixing interferometers are designed with a single lasing mode laser. However, different conditions like temperature, operating current of laser and, length of external cavity can cause mode hopping phenomena, which results in the generation of multi-modal SMI signals, leading to measurement errors. Such type of signals can be detected and corrected by adjusting the aforementioned conditions which are responsible for mode-hopping. The total number of transitions of an SMI signal is a statistical feature, which represents the non-linearities in the signal and can be a good distinguishing feature between mono- and multi-modal SMI signals. This paper purposes a technique based on transitions of SMI signal to detect multi-modality, even in the presence of noise and speckle. Furthermore, the proposed technique is able to auto-compute the values of the distinguishing feature. The proposed technique detects multi-modality with a success rate of 97%.