Efficient vibrotactile coding should rely on perceptual models to reduce perceived quality loss caused by signal compression. So far, single-point models have been proposed that take frequency masking into account, leaving spatial masking between multiple points of interaction underexplored. This paper investigates spatial masking effects in vibrotactile perception and extends a frequency-masking psychohaptic model by incorporating a spatial attenuation factor. A psychophysical experiment is conducted that measures masking across finger locations and frequencies, revealing that actuator positions significantly influence masking strength, while frequency dependence follows trends observed in same-site masking. No interaction between position and frequency was observed. Based on these findings, we derived a new multi-point spatial psychohaptic model. A pilot study is conducted to validate the model, demonstrating its ability to achieve stronger signal compression before perceptual degradation becomes noticeable. The results highlight the potential of spatial masking models for optimizing multi-point vibrotactile codecs. Future work will investigate distance-dependent attenuation, broader spatial configurations, and applications beyond the tested scenarios.