Human memory did not evolve in a vacuum. Rather, memory systems were shaped by natural selection to solve fitness-relevant problems, such as the efficient location of valuable calorie-dense foods. We explored whether a high-calorie bias in human spatial memory exists, and can be reliably observed across sensory modalities, participant populations, and experimental paradigms. In two lab studies with distinct samples of 88 participants, individuals had to re-locate foods on a map in a computer-based spatial memory task using visual (Study 1) or olfactory (Study 2) cues that signaled (sweet and savory) high- and low-calorie items. Individuals consistently displayed lower pointing errors (i.e. enhanced memory) for locations of high-calorie versus low-calorie foods (Study 1: 121.6 vs 137.8 pixels; F(1,4049) = 8.25, p =.004; Study 2: 118.3 vs 152.7 pixels; F(1,1240) = 18.43, p <.001) – regardless of hedonic evaluations, familiarity with foods, or encoding time. In a final field-based experiment, we covertly tested the food spatial memory of 512 individuals who navigated a maze-like food setting, and either ate (high- and low-calorie) food products (N=258; multisensory environment [ME]), or smelled corresponding food odors (N=254; olfactory environment [OE]), at eight spatially dispersed pillars. Individuals incidentally learned and more frequently re-located high-calorie stimuli to correct pillar locations (ME: 0.63 vs 0.57 proportion correct relocations, χ² (1) = 9.35, p =.002; OE: 0.36 vs 0.30, χ² (1) = 6.88, p =.009) – while controlling for explicit liking, desirability, and familiarity with foods. The high-calorie spatial memory bias was also equally expressed in both sensory environments – even where solely odor information was available. Taken together, human spatial cognition seems to be optimized for ancestral priorities of energy-efficient foraging. This research was financially supported by the Edema-Steernberg Foundation and the Netherlands Brain Foundation.