First, when you are doing bioimaging experiments, you want to keep your cells alive. Thus you don't want highly energetic wavelengths such as UV or violet which can heat your cells, damage DNA, produce ROS ...
Second you want to have a good penetration of the incident light in your sample, with the lesser refraction, so you want the longer wavelength. But as the Ernst Abbe's axiom says that the limit resolution is a function of the emission wavelength (roughly lambda/2), if you want a reasonable resolution you will try not to go too far in the red and infrared.
Third the probes you are looking for will tell you what wavelengths you need. For example if you use GFP, SYBR Green, FITC, you will need a blue laser (488nm), if you use mCherry, TRITC, PI, you will need a green laser (514 or 532 nm). If you use NIR probes, you will need 561 or 633 nm laser. If you are rich enough, you can buy a tunable IR femtoseond pulsed laser and do two-photon fluorescence microscopy with a single laser, but it is horribly expensive, loudy, electricity consuming, heat producing and the maintenance is really difficult and costy.