The main question is from my point of view, how do you define "quality of SEM images"? You don't say anything about the signal you are using, the used magnification, the applied energy of primary electrons, the beam current, beam size (resolution), the type of SEM (cathode etc., immersion lens etc), attached equipment  like other detectors or anything else, the material you are investigating. The number of factors is quite big, and for practically each of them different options exist in order to improve the derived image using counted electrons. The entire imaging in SEM is always very critical to discuss: Were the conditions during scanning properly chosen so that the derived image really shows me a signal which correlates with the sample surface, or is it more an interference of side effects? Since electron imaging is not really comparable with a light microscope I hardly recommend to keep the image formation in mind! 

Carbon/gold/chromium coating is needed for non-conductive materials

use a sputter coating machine to coat your samples with a conductive coating,i.e., gold, carbon, in a vacuum chamber .   The sputter coating will reduce the electrostatic charge   of  a non-conductive sample in the SEM.   As a result, it  will improve  the images.   The principle of SEM works based on detection of the inelastic scattering of ( secondary) electron of a conductive surface.   Therefore, a conductive surface of the examine sample is must for quality SEM images.      

The main question is from my point of view, how do you define "quality of SEM images"? You don't say anything about the signal you are using, the used magnification, the applied energy of primary electrons, the beam current, beam size (resolution), the type of SEM (cathode etc., immersion lens etc), attached equipment  like other detectors or anything else, the material you are investigating. The number of factors is quite big, and for practically each of them different options exist in order to improve the derived image using counted electrons. The entire imaging in SEM is always very critical to discuss: Were the conditions during scanning properly chosen so that the derived image really shows me a signal which correlates with the sample surface, or is it more an interference of side effects? Since electron imaging is not really comparable with a light microscope I hardly recommend to keep the image formation in mind! 

You really need to tell us more about the equipment you are using, the material you are trying to image, what sample preparation methods you have already applied and what operating parameters you are currently using. It would also be helpful to post an image that you have produced that you are unhappy with and an image showing the sort of quality you are hoping for. (Occasionally a student will come to me and show me an image in a published paper and say I want to do this - but the specification of our £300K EM are not up to the level required and it would need a spend of £1M in order to replicate the results.) Also the level of experience of the operator has a lot of control over the final image.

Image 'quality' is difficult to define as mentioned by Gert, as in certain circumstances I will deliberately reduce the resolution of the image in order to improve the apparent quality! This can happen at lower magnifications where if your resolution settings are too high then noise takes over and reduces the quality. At higher magnifications you need to do the opposite in order to improve 'quality'. So as you see without further information on what you are trying to achieve it is difficult to help you further. 

However, very generally speaking, if you are dealing with non-conductive materials the options are - make it conductive, coat as mentioned by Khaled, OR work in a low vacuum or variable pressure or environmental SEM this will negate the charge, OR work in a FEG-SEM at very low landing voltages, using beam deceleration techniques. 

Can you attach your images? 

Talking about abstract quality is useless, too many variables.

The sputter coating is required for non conductive materials. Also check Back scatter image in SEM .

Actually, the previous answers are all correct, and the best solution depends on your type of microscope and detectors you have.

For an older W-filament SEM: sputter coating (ex: carbon, Au)

if you have variable pressure/extended pressure options then without coating you might get good results. some newer SEMs have options to spray N2 gas locally to.

if you have a FEG-SEM:

-lower WD, use 1-2 kV, and use in-lens detectors

- try using gentle-beam options (apply voltage to sample to reduce effective kV)

also try backscatter imaging, if you have a detector with 4 (or more) quadrants for detection, subtraction signal from one quadrant can give you pseudo-secondary electron images

- for beam-sensitive materials faster scan rates might be useful, you can integrate several frames gathered extremely fast

- non-conductive specimens can cause "drift"; which is another source for low-Q images; if the drift is occuring regularly, it can be controlled (there are even commercial solutions)

Happy imaging !

Based from my experiment, 1st time using SEM always desperate. but by the time you used it, you will comfortable to used it and you will get the good image. If the coating is not really helpful, tray to double it or make the coating time little bit longer.

you can make sure, if your SEM is old type (U have been used old JEOL SEM), try to put the SEM in some cold room (I mean really cold room temperature in AC). you will get better image.

Higher KV will give better resolution and crisper images.

By optimizing the aperture, working distance, sample geometry and energy input may provide better images....It's a practice to achieve...

For non conductive materials, it should be ion sputtered for at least 10 seconds to improve the charging thus provides good electron scattering.

Sorry for my ignorance, I am new using an SEM, I have many questions to ask, but this time I will ask only one question. What are the steps to follow to obtain an image with secondary electrons?

Follow a manual to your SEM. If you have more specific questions, ask them in separate thread, do not forget to mention what specimens and for what purposes you are using.