In this paper, we focus on analyzing multi-location charging behaviors of plug-in electric vehicles (PEVs) under the Time-of-Use (TOU) pricing scheme. Trip chain model incorporating the information on start time, end time, driving distance, start location and end location of each trip is used to depict the sequence of daily driving missions. Statistical distributions of travel patterns are fitted from a real-world driving dataset. Then a method is developed to generate the complete trip chain of each individual PEV. Simulation results show that due to the TOU price, people prefer the overnight charging at home, although workplace charging can support PEVs with smaller battery capacities and long-distance commute trips. Furthermore, with a large enough battery capacity, charging at workplaces will not be necessary if vehicle-to-grid (V2G) technology is not considered. Although high charging power accelerates PEV's charging processes during price-valley periods, it imposes higher requirements on the charging facilities and distribution grid.