We present results from a study on how the regolith surface affects the measured soft X-ray spectra from atmosphereless planetary surfaces as a function of viewing geometry. The analysis of the spectra obtained by orbiting spacecraft uses relative fluorescent elemental line intensities (e.g., Ca/Si, Fe/Si) as the first-order approach to determine the mineralogy at the target surface. It has for some time now been a subject of discussion how much the analysis is affected by the viewing geometry. We have performed laboratory measurements to simulate the observations from an orbiting spacecraft and can confirm that the fluorescent peaks at higher energies are enhanced at larger phase angles (i.e., the angle between the light source and the observer as seen from the target surface) thus changing the measured relative intensities. The measurements have been performed with a high-resolution X-ray spectrometer on olivine basalt samples with two different grain sizes over the phase-angle range of 7-51 degrees. Complementary to the laboratory measurements we have also developed a numerical ray-tracing Monte-Carlo simulation code. Our code simulates the fluorescent signal induced by realistic solar-like X-ray flux from a regolith consisting of spherical particles with different physical properties. The simulations and their interpretation in light of the empirical studies are discussed. We discuss the importance of the present studies for the future spacecraft missions carrying X-ray spectrometers, such as Selene and Chandrayaan for the lunar exploration and Messenger and BepiColombo for hermean exploration. BepiColombo will carry a focusing X-ray imaging spectrometer called MIXS-T, first of its kind on space missions, which has a very high spatial resolution. The need for better understanding of how the viewing geometry changes the signal is crucial for that mission. We conclude by presenting future plans for more detailed studies.