There are rich data in the manufacturing information systems, but they are not utilized in an effective way. The establishement of assembly data mining platform is to take advantage of the data to improve assembly process. Assembly factors and assembly data mining templates are the core tables in the assembly database. The methods of dealing with data vacancies and noise points are pre-set in the data mining templates. Python is used as the data mining engine by script customization and algorithm library encapsulation. Firstly, Python algorithm scripts are customized when programing the platform. Then the platform generates the execution script according to the user operation. Finally, the main program of Python executes the generated script and returns the results. Also, Mlpy is applied to make corresponding algorithm processing module. The functions are pre-compiled so that the assembly technicians without knowledge of data mining can utilize the assembly data to predict the assembly performance and analyse the assembly process.