Path planning algorithms for intelligent agents are designed to plan the behavior trajectory of an agent so that it can safely and efficiently reach the target point from the starting point without colliding with obstacles. Currently, path planning algorithms have been widely applied in various critical cyber-physical systems. Therefore, it is essential that the path planning algorithms be tested before being put into use to evaluate whether their performance can meet the requirements. However, the distribution patterns of threat obstacles in the task space, which are the inputs of the path planning algorithm, are complex and diverse. Moreover, a relatively high operational cost is usually required when the path planning algorithm plans a path for each test case. To improve the testing efficiency of the path planning algorithms, this study adapts the concept of dynamic random testing into path planning algorithms and proposes the dynamic random testing approach for intelligent agent path planning algorithms (DRT-PP). Specifically, DRT-PP discretely divides the path planning task space and introduces the threat generation probability within each sub-region, thus constructing the test profile. This test profile can be used as a testing strategy in the process of test case generation. Furthermore, the test profile is dynamically adjusted by DRT-PP during the testing process to make it gradually optimized, thereby improving the testing efficiency. Experimental results show that, compared with random testing and adaptive random testing, the DRT-PP approach can not only ensure the diversity of test cases but also generate more test cases that can expose the performance defects of the tested algorithm.