You begin with choosing a graphics API. I suggest OpenGL 3.x, as it is cross-platform and has really great support. Next, you have to learn matrix algebra. Then you have to use matrix algebra to calculate the 3D projection. Once you have that set up, it's easy. You can use the GLEW library to gain access to OpenGL functions.
For the voxels, you can use a simple 3D array to store the data and use a noise generation algorithm to generate their points. After your 3D projection is ready, you can just pass the coordinates directly to the graphics card.
If you don't understand what I'm saying or don't know how to apply it, this is proof that it's hard to make a game in C++ if you don't know 100% what you're doing. It's one of the hardest languages. Concepts are much more lower level.