Black pepper (Piper nigrum L.) is a crucial spice crop that is very vulnerable to various foliar diseases that highly affect the yield and quality of the harvest. In the current study, it is proposed that a deep learning model based on Vision Transformer (ViT) can be used to automatically identify the three types of leaf diseases of the Black Pepper: Footrot, Pollu Disease, and Slow Decline. ViT model is effective in utilizing self-attention mechanisms to extract small visual details, the variations of texture, and overall morphological patterns peculiar to each type of disease using raw leaf images directly. A hand-selected set of Black Pepper leaf image was used and preprocessed and augmented to make the model resistant. Through experimentation, the ViT model proposed has been shown to attain a 99.33% classification accuracy, which is higher than the performance of the traditional convolutional models in terms of precision and generalization. These findings validate that transformer-based architectures may be efficient instruments of accurate and trustworthy disease detection in Black Pepper leaves, which can be used to detect diseases at the earliest, as well as manage crops sustainably.