The direct-start method without requiring a starter is a viable and cost-effective solution for generating a frequent and efficient restarting process when start–stop technology is used on a gasoline direct-injection engine. During the direct-start process, the first-cycle combustion characteristics play a key role in determining whether the start mode is successful or not, because the start energy is wholly derived from the in-cylinder combustion without the support of a starter motor on the gasoline direct-injection engine. However, the first-cycle fuel–air mixing and combustion characteristics during the direct-start process of a gasoline direct-injection engine are not fully understood. In this work, the influences of the injection parameters, including the delay between the injection and ignition, the excess air ratio for the single-injection strategy, the delay between the first injection and the second injection for the dual-injection strategy and the ratio of the fuel mass in the first injection to the fuel mass in the second injection for the dual-injection strategy, on the combustion pressure, the heat release rate, the accumulated heat release and the indicated work were investigated experimentally by cycle-by-cycle analysis. The results show that the optimal delay between the injection and ignition of the single-injection strategy was 200 ms as a longer delay or a shorter delay can result in a reduction in the heat released rate, the indicated work and the firing boundary. A shorter delay with the optimal injected fuel mass tended to be more beneficial to the accumulated heat released, the indicated work and the crankshaft speed. Furthermore, with increasing delay and increasing fuel ratio of the fuel mass in the first injection to the fuel mass in the second injection for the dual-injection strategy, the heat release rate, the accumulated heat release and the indicated work first increased and then decreased. The optimum delay was 10 ms and the ratios of the fuel mass in the first injection to the fuel mass in the second injection were 4/1 and 5/1 respectively under the test conditions. Additionally, the dual-injection strategy with an optimized control parameter produced a higher heat release and higher indicated work than the single-injection strategy did.