The objective of this dissertation is to investigate the performance of the recently proposed MIMO technology in real indoor environments based on channel measurements centered at 5.8 GHz. First, a MIMO channel measurement system is implemented based on the virtual antenna array infrastructure. This measurement testbed can acquire the wideband channel matrices of MIMO systems with arbitrary array geometries. The measurement system structure and measurement procedure are described in detail in the first part. The second part is about MIMO channel modeling. Two novel number-of-sources detection algorithms, which are more robust and suitable for practical applications than traditional methods, are proposed. The MIMO path parameters, including delay, DOA, and DOD are estimated from measured data by several estimation schemes based on the ESPRIT algorithm. The accuracies of these estimation schemes are evaluated in terms of the estimation error between the capacities of the directly measured and the reconstructed channels. Moreover, based on ray tracing and measurement results, the spherical wave model is suggested to replace conventional plane wave model in order to prevent the capacity underestimation of short-range MIMO channels. An important observation is that short-range MIMO can achieve full capacity in free space channel. A threshold distance is derived to determine whether the spherical wave model is necessary. In the final part, measurements conducted in the Residential Laboratory are used to investigate the impact of element spacing, LOS, interference, spatial correlation between the interfering and data links, and stream control. A capacity enhancement scheme, which improves the performance by adapting the element locations, is implemented using our measurement system. Finally, the performances of beam selection and antenna selection in combination with MIMO technologies are compared in both narrowband and wideband channels.