The invention discloses an evaluation method and device of a vehicle message data model. The method includes the following steps: collecting vehicle message data of a predetermined time length, and dividing the vehicle message data into different data sets; for data in each data set, setting a part of the data as training data, setting a part of the data as test data, and setting the rest of the data as verification data, wherein part of the data in the test data is modified as simulated abnormal data; inputting the training data into a long-short time memory network to establish a message prediction model; testing the test data on the basis of the message prediction model to obtain the detection rate and the false alarm rate of the simulated anomal data; evaluating the message predictionmodel on the basis of the detection rate and the false alarm rate. Compared with prior art, the evaluation method and device have the advantage of reducing the workload of vehicle abnormal data detection.