It is challenging for real-time database systems (RTDBs) to support transaction timeliness while preserving data freshness under unpredictable workloads. So this paper proposed a scheme for data and transaction quality management to solve this problem. A definition of data quality considering data characteristic in both time and value domain was given. Based on the definition, a data quality adjustment method was presented which adjusts update periods and dynamic similarity degree lower bounds. This scheme uses the adjustment method, feedback control and admission control to support Qos targets. Experiments have shown that the proposed scheme outperforms the existing schemes.