We consider facility layout problems with shape and position constraints for each facility. We use a simulated annealing (SA) algorithm for finding a layout in which sum of rectilinear distances weighted by flow amounts between facilities is minimized while satisfying the shape and position constraints for facilities. A solution of SA is encoded as a matrix of numbers. We convert the matrix into a layout by partitioning a floor plan into a set of rectangular blocks while satisfying areas of facilities. In the partition procedure, we suggest three methods to minimize the number of facilities which violate shape constraints. Computational results showed that the proposed algorithm gave better solution than existing algorithms, especially for problems which have tight shape constraints.