Currently， gradient game algorithm is commonly used for indoor virtual space positioning of full-scale residential buildings， but it is susceptible to the influence of the standard deviation of indoor virtual space scanning numbers， resulting in a large deviation distance between the positioning results and actual values， and low positioning accuracy. Therefore， a full-scale indoor virtual space positioning algorithm based on color flow model is proposed for residential buildings. Based on the relationship between indoor virtual space regionalization and indoor positioning， a color flow model is used to encode indoor image features， achieve regionalization processing of indoor virtual space， and construct a distribution function of indoor virtual space. Combined with the fitting quantification value of spatial scanning data， the indoor virtual space coordinates are transformed. The standard deviation of indoor virtual space scanning number is calculated based on the error propagation formula. By mapping the key feature points of indoor virtual space within the scale space and combining them with the beacon sequence of key points， the position coordinates of indoor virtual space are finally obtained to complete spatial positioning. The application results of the example show that the spatial positioning offset distance obtained by the proposed method is small， and the positioning accuracy is high.