"SCI-indexed" means the journal appears on the Scientific Citation Index; ou can find the list online.
Generally speaking, most legitimate science journals are on this list and many not-so-legitimate journals are not on it. There are exceptions, but this (along with other indexes and lists) is a useful way to judge if a journal is legitimate and has a good reputation. Although to be honest the best advice is just to only publish in journals that you are already familiar with somehow (i.e., journals that have published the papers you cite, or journals that you have reviewed for and seem to be serious, etc.).
The role of SCI is strongly field dependent. I'm afraid in mathematics, for one, there is a number of quite reasonable journals which are not SCI-indexed.
However, by and large I tend to agree with Stephen that for a journal being SCI-indexed is one of the indicators (however, neither necessary nor sufficient condition!) that a journal is reasonable.
The Science Citation Index (SCI) is a citation index originally produced by the Institute for Scientific Information (ISI) and created by Eugene Garfield. It was officially launched in 1964. It is now owned by Thomson Reuters.
You can refer to the links for the list of journals.
SCI is a subset of SCIE, covering the most highly cited & high impact factor Journals in each category. SCIE is an interdisciplinary index, covering more than 8,500 notable and significant journals, across 150 disciplines.