Wind energy is the front-runner of renewable energy sources. Its installed capacity in the U.S. has increased more than 20 folds in the past 16 years, from 4.2 GW in 2001 to 82 GW 2017. US Department of Energy envisions that wind will generate 20% of the nation’s electricity by 2030 and 35% by 2050. China has been the largest wind energy provider worldwide since 2010 and its installed capacity in 2016 was 149 GW. Chinese government pledges that renewable sources, wind included, produce 15% of the nation’s electricity by 2020. The ever-changing wind exerts a non-stationary and non-steady load on wind turbine drive train, causing wind turbines to deteriorate faster than other turbine machineries, and other harsh environmental conditions such as icing and lighting add to the low reliability of wind turbines. Low reliability drives up the cost of operations and maintenance and becomes one major obstacle towards wind energy’s market competiveness and viability without government subsidy. In this talk, we will discuss the reliability issue in wind energy, approaches to assess the system-level performance for a wind turbine generator and strategies to countermeasure the decline in a turbine’s power production capability, as well as the data science relevance to addressing research challenges in wind energy applications. This talk was previously presented on July 6, 2018 at the 8th International Workshop on Reliability Technology and Quality Science (RTQS-2018) at the Chinese Academy of Sciences in Beijing, China.