Abstract: To solve the problems of unclear vehicle operation speed pattern, single safety driving speed limit, low production efficiency, incomplete prediction model and insufficient data set, a multi-source data fusion analysis model considering multiple factors and time characteristics is proposed to analyze vehicle operation characteristics and speed prediction. Firstly, the vehicle speed meteorological data, UAV images and production plan drawings data are collected in an open-pit mine for two years, and the data are cleaned, coordinate transformed, and the data are-dimensional integrated. The mathematical methods such as cluster analysis and correlation analysis are used to obtain the operation speed characteristics of the vehicle under different meteorological road morphology. Secondly, the road, operation period, meteorological and other factors are integrated to deduce the multi-factor open-pit mine vehicle speed prediction model under the condition of no precipitation, light rain, moderate, heavy rain and snow at different moments, and the average relative error, average absolute error and root mean square error of vehicle speed are used as the evaluation indexes of the speed prediction. Finally, the vehicle safety driving speed model is constructed according to the road friction coefficient, the driver’s reaction time, the vehicle braking distance and other parameters, and the safety driving speed under different space-time scenes is obtained. The experimental results show that: ① the vehicle speed is in a high operation state from 7 am to 1 am, the vehicle speed is quite different under different conditions, the change range is between (-46% ~ 11%), the speed increase range at different times on same day is between (-20% ~ 5%), and the vehicle transportation efficiency will present different efficiency at different times; ② the vehicle operation speed is negatively correlated the precipitation amount, and it is strongly correlated with the road structure and time distribution; ③ through the analysis of the constructed model, it can be seen that the average relative error the model is within 2%, the average absolute error is controlled within 0.4 km/H, and the root mean square error is less than 2 km/, the prediction result is better; ④ taking the safety driving speed and the minimum value of the predicted speed as the safety driving threshold, and obtaining the safety driving speed threshold of road morphology at different moments under different weather conditions. The research results can improve the accuracy of speed prediction in different sections of the open-pit mine, and provide support for the correlation the operation law of vehicle speed and production organization, and real-time dispatching. It is of great significance to the production space-time steady connection of the open-pit.