I've received a lot of notes from people who are interested in getting into concept art and/or landscapes and I always struggle to come up with advice that doesn't seem awfully confusing. I can't say I'm a veteran in this industry, but since people keep asking me and I do have experience, here's my 2 cents (PSA: by no means a comprehensive guide or the absolute truth). I am very humble when it comes to learning this stuff and I never know enough myself, but these are some of my thoughts that came up when a student sent me a note today;
My best advice is to climb to the tree from the root and accept that there's no shortcut.
Concept art is a very loose term. It can mean anything from sketching characters, enemies, vehicles and objects to doing detailed plans of level visual design or landscapes. Game concept art can also serve a lot of purposes. Most of the time the most crucial purpose of concept art is to communicate (a) game designers' idea of how a game looks and feels. So for that you need to know how to draw pretty much anything. What is always important is that you can sketch out a piece that communicates to the viewer how a certain place feels to the audience; what are the scale and proportions like, communicating through colors if there is a feeling of peace, serenity or action, using lighting to give focus to important elements... The list could go on forever. There is no bottom line. Concept art is hard to 'make up'. You need to have at least an imagined 'customer' (this can be yourself) that has certain requirements, a bit of a backstory, an idea that they want illustrated. The best you can do for portfolio is ask maybe a game designer or a writer friend to ask for an illustration for something they would like to see, something that is a visual representation of a concrete idea they have in their head; maybe a scene (hero enters a foreboding mansion to avenge his dead father, it is a stormy night, he is afraid but determined) or a game idea (a platformer drawing inspiration from Indiana Jones where your goal is to collect shards of crystal to complete a weapon to save humanity from an evil space lord, the level takes place in a sci-fi lab) just to name a few.
As for landscapes;
Learning everything from color theory and lighting and perspective gives you the necessary information that you need in order to build on top of that. If you just get into landscapes and skip the theory, it might feel quicker but it will stab you in the back because that stuff is something you will have to go back to learn when you realise you can't advance without having learnt it. For me the best way to do good landscape art is to be genuinely interested in what you are drawing. I draw nice trees because I'm honestly interested in how they look. Observe real life and draw. I use a lot of time just looking at stuff when walking my dog. You don't even need to draw it. Just look at whatever is around you in the city or nature and try to figure out that if you were to replicate that on canvas, what colour everything is. Ie. Is the treebark darker than the stones, what is the darkest colour if this were a picture, what colour is the light, how does it affect the different textures around you?
My main piece of advice is to draw from nature and photographs. Don't just mimic and replicate pictures and copy, think for yourself. Sometimes its more important to communicate the feeling in the picture and how you see the landscape than getting stuck in the details (like trying to draw something in a photorealistic way just for the sake of realism). Getting stuck in the details and focusing on execution rather than message and idea often ruins pictures and they become uninteresting to look at.
Those are my thoughts, hope it helps.