Spatial clustering, which groups similar objects based on their distance, connectivity, or their relative density in space, is an important component of spatial data mining. The problem of detecting clusters of points belonging to a spatial point process arises in many applications. One of the challenges in spatial clustering is to find clusters under various cluster number, object distribution as well as multi-density. In this paper, we propose GDCIC, a Grid-based Density- Confidence-Interval Clustering algorithm for multidensity in large spatial database. By using the technique of confidence limits of the density confidence interval, accurate density estimation in local areas can be produced to form local density thresholds. Local dense areas are distinguished from sparse areas or outliers with the help of these thresholds. An optional procedure is included in GDCIC to optimize the clustering result. The experimental studies on both synthetic and real datasets show its high accuracy and performance over existing algorithms.