I investigated the ability of a slime mould Physarum polycephalum to find food in a complex environment and compared it with random search. For this, I created a maze with the starting point and three different routes leading to the end point. At these points, we placed oat flakes as the source of energy and award, respectively. I examined whether P. polycephalum can reach the end point successfully (via the shortest route) and efficiently (minimum distance and time in suboptimal paths). Analysis showed that Physarum had a 89% success rate which outperformed 46% random search by a computer algorithm. However, its efficiency in time (34%) and energy use (54%) was limited. These show that P. polycephalum outperforms random search suggesting it may be a candidate for biological computing.