In cyber-physical system (CPS), sensor-based communication components are widely deployed to monitor real-time physical phenomena and send the information to network to enable control of entities in physical world. The freshness of data is of critical importance in such systems. Age of information (AoI) is a performance metric which measures data freshness at the destination since the last sampling at the data source. In this letter, we consider a multihop CPS with multiple data flows and investigate an age minimization problem by jointly optimizing data sampling at the sources and link scheduling along the path of each data flow. The formulated problem falls in the form of an integer nonlinear program. To solve the problem efficiently, we propose a light-weight scheduling algorithm. Compared with traditional delay-based scheduling scheme, simulation results show that AoI-based scheduling scheme performs better in delivering information in a timely manner. Moreover, we demonstrate that by jointly optimizing link scheduling and sampling time, network-level data freshness can be further improved.