With the sweep of the era of “Software Defined Vehicle”, software related issues account for over 70% of automotive problems. Software testing includes simulation testing, field testing, and road testing. With the continuous compression of the entire vehicle development cycle, relying solely on real vehicles to discover software problems is no longer sufficient to meet the constantly compressed entire vehicle development cycle. Simulation testing can detect software issues in advance to meet the increasing software quality requirements of self-developed controllers. At present, industry software development and testing both include simulation testing, and HIL testing has become an indispensable software testing method for host factory software development. At present, there are still the following problems in hardware in the loop simulation testing: most HIL testing in the industry is at the component level, and there is very little interaction testing between systems, and even whole vehicle level HIL; This will lead to incomplete testing; The test cases lack correlation. HIL testing has not formed a systematic technical system; Not establishing a complete database; Failure to analyze and summarize based on the database, and identify the occurrence rate of problems; HIL testing is limited to laboratory testing and cannot be conducted remotely, which affects testing efficiency. Based on the above issues, this study constructed a cloud platform to achieve cloud based testing; Accumulated databases: requirement library, use case library, scenario library, model library, problem library, etc.; Conquered the HIL testing method; A multi-level integrated HIL verification system has been formed. Identify software issues in advance to ensure smooth trial production and installation of vehicles, and improve software quality after vehicle SOP.