Cognitive radio (CR) is a promising technology for overcoming the lack of available communication bands. In CR technology, spectrum sensing is an important issue, which has recently been extensively studied. We provide a solution to the spectrum-sensing problem for multiple cognitive terminals (CTs) that takes into account the difference among CTs with respect to the probabilities of a false detection and a missed detection. We optimize the spectrum-sensing performance by differentiating the number of spectrum-sensing operations that each CT performs. This has not previously been proposed in the literature. Moreover, we use a simultaneous spectrum-sensing and data transmission scheme that utilizes multiple-input-multiple-output (MIMO) antenna technology. As a result, the degradation of quality of service (QoS) that is caused by spectrum sensing can be reduced, and the throughput of CR systems can be increased, while maintaining the accuracy of spectrum sensing. Through performance analysis, we show that our proposed scheme can achieve the desired levels of performance with respect to the probabilities of a false detection and a missed detection and improve the performance of the CR systems with respect to throughput and delay.