Once upon a time, I dispensed lots of advice to fellow writers. I still do, one on one. But I stopped blogging about the writing craft when it became clear that the people who are interested in that sort of thing are other writers, not readers. The question comes up often, though, so I'm going to take a crack at summarizing all my best advice. Here goes:
1. First, read good novels
You don't learn to speak a language fluently in the classroom. While you can pick up the rules and rudiments there, to become truly proficient, you have to listen to people who speak the language well. The same is true when it comes to writing fiction. The single best thing you can do (and the single most enjoyable) is to read good novels.
Reading bad novels will teach you plenty, too, but it won't inspire you to excell. When you read a bad novel, you end up thinking, "I could do that." You begin to write with the idea, not of doing your best, but of improving on the crap that passes too often for fiction. This is like watching a bunch of awkward kids at Little League practice and deciding to go for the Majors.
So what's a good novel? It's the one you read and wonder afterward, "How did he do that?" Pose that question and stick with it until you come up with an answer. That'll teach you better than anything how to write a novel.
2. If you must read a how-to book, make it Stephen Koch's The Modern Library Writer's Workshop
I'm not saying it's the only, or even the best how-to book out there, just that it's the best summary of the essentials, and includes an annotated bibliography in back that will guide you through the rest. I spent years in an excellent MFA program, and I don't think I learned anything of value that isn't covered by Koch to one extent or another. Think of this book as orientation.
The thing that prevents most novels from being written isn't poor writing, it's poor planning. Back in 2007, I wrote a series of posts called "Planning a Novel" outlining my own method. Think of the planning process as the time you spend (a) learning your story and (b) figuring out how to tell it. A lot of the work you do during this period won't end up in the finished manuscript. That doesn't mean it's wasted. The goal of planning isn't to map everything out in advance; it's to know your material so well that you can take it in whatever direction the story wants to go.
Here at Crime Genre, I occasionally break my own rule and write about craft. Before this, I wrote almost entirely about craft at a blog called Write About Now, which went dormant in 2010. (Prior to that, I wrote a dedicated blog for writers called Notes on Craft, which isn't online anymore. Not to worry: some of the best material was reposted at Write About Now.)
5. One last thing: Don't put too much stock in the "rules"
Grammar is one thing, craft is another. Now that there's an industry that exists to educate (and profit from) aspiring novelists, an unparalleled volume of how-to guides are being written, most of them based not on fresh experience but on a regurgitation of earlier how-to guides. Often, these sources will inform you about the things you must do -- or not do -- in order to sell your work.
You won't sell your novel by following a set of rules any more than you'll fail to sell it by breaking them. If only it were that simple. The fact is, many of the rules are either helpful observations that have been misunderstood and misapplied over time (like "show, don't tell"), or particulars drawn from one writer's success then universalized (such as the notion that if you write in the style of James Patterson you will sell a similar number of books). Try to tune this stuff out.