To improve the tracking speed of maximum power point (MPP) of photovoltaic (PV) system, some maximum power point tracking (MPPT) algorithms or their improved algorithms have been studied in depth. However, these algorithms are usually the improvement for a certain MPPT method, now there still exists no common method for improving the rapidity of the MPP tracking. To solve this question, in this paper, the conceptions of load-variable-weather-parameter (RVWP) optimization interval and variable-weather-parameter (VWP) optimization interval are proposed. Meanwhile, on the basis of them, a VWP optimization strategy, which can optimize the tracking speed of most MPPT algorithms, is proposed. In this strategy, the control signal of PV system has been quickly contracted within RVWP interval or VWP interval before a MPPT method which will implement the MPP tracking is operated. Therefore, the essence of this strategy is that the MPP tracking speed is increased by contracting the optimizing range in other words, the global optimization is converted into local optimization. Finally, in this paper, the conventional perturb and observe (P&O) method is selected as the optimized object to study the MPP tracking speed under constant load, changing load, constant weather and changing weather conditions. The results of simulated experiments show that the MPP tracking speed can be greatly optimized by VWP optimization strategy with RVWP interval in any case and can be optimized by VWP optimization strategy with VWP interval in most cases, and that the output power stability can be also optimized by proposed optimization strategy for PV system with output power oscillating.