I'm just curious, but I've seem some questions with lots of votes. But how can these moderators know that the votes on the questions are real? There has to be some detections that goes with it. So, my question is that is there a system to detect if a user has created multiple accounts just for vote their own questions and earn fake reputations and harm other users?
I've seem some questions with lots of votes
There is normally a strong correlation between number of views and number of votes. Questions and answers with 10,000 views are going to have lots more votes than ones with 50 views.
But how can these moderators know that the votes on the questions are real?
We, the amateur unpaid moderators, have tools. The SE employees above us have much better tools (and more time, it is after all what they are paid for). Sometimes we spot someone operating multiple accounts. Sometimes we are alerted by an alert user flagging something suspicious.
Sometimes (probably the majority of times) SE employees detect and delete such accounts, often across several platforms, and the first we low level mods know is that we get a message like "you lost 20 points because we just deleted an account which also upvoted 2 of your questions/answers".
If the multiple user accounts are low-reputation (<2000) we can take action up to and including deleting accounts. For higher reputation accounts we simply report up to the paid employee level where they have powers to investigate for likely cross platform multiple accounts. After all, if you were successfully operating multiple accounts why would you limit it to Chess SE?