The traditional methods for predicting the effectiveness of public opinion dissemination have average performance in terms of prediction accuracy, therefore, in order to improve the prediction accuracy of public opinion dissemination effect in emergencies, this study proposes a new intelligent prediction method using Bi-LSTM. First, construct the relationship diagram of public opinion elements of emergencies, and analyze the indicators of various influencing factors; Then, using Python tools and the Term Frequency Inverse Document Frequency (TF-IDF) method, extract the characteristics of public opinion propagation in sudden events, mainly including temporal and textual features; Finally, based on the extracted features, the Bi-LSTM model is used to accurately predict the public opinion dissemination effect of emergencies. Through testing, it can be seen that after applying this method, the error between the predicted and actual results of the public opinion dissemination effect of sudden events is small, indicating that the prediction accuracy of this method has been significantly improved.