The traffic patterns of Spiking Neural Networks (SNNs) are highly varying, which will cause elevated traffic hotspots for the Network-on-Chip (NoC). How to predict the occurrence of hotspots is still one of the most challenging issues for NoC design. It is the first work that presents an attempt toward utilizing a Liquid State Machine (LSM) in predicting the formation of the hotspot. We extract the most critical information which affects the emergence of hotspots. We also adopt the heuristic algorithm to search the hyperparameter of the LSM. Results indicate that the predictor can forecast hotspot formation with high accuracy