Crystalline silicates are observed in outer cold regions of a protoplanetary disk although they would be formed by annealing of amorphous precursors in the inner regions of the disk. Although some studies showed that small crystalline silicate dust could be transported to the disk outer region, the crystallinity of dust considering size growth is not well-investigated. Arakawa et al. showed large crystalline silicates could be carried in an evolving disk due to disk wind. However, they did not consider the difference in stickiness between ice and silicate. On the other hand, although Okamoto & Ida showed the crystalline abundance increase due to the higher stickiness of ice than silicate and decaying icy pebble flux, they did not consider the size growth of silicate dust.We performed a 3D Monte Carlo simulation of advection/diffusion of silicate particles in an expanding disk considering dust growth and evaporation and condensation of water. We assumed the stickiness of silicate is less than ice, and when particles collide faster than velocity for fragmentation, they are fragmented. We set the “annealing line” by T = 1000 K and assume all amorphous precursors are annealed inside the snow line. At first, icy pebbles including amorphous precursors grow up fast and drift inward. Then, icy pebbles pile up at the snow line. As a result, most silicate dust does not diffuse over the snow line and more amorphous silicates are annealed. After that, due to the inward movement of the snow line, some crystalline silicates can diffuse over the snow line. Therefore, the crystallinity of silicates around snow line is raised up to ~ 90% for 1 Myr. Finally, since mass flux of dust inside snow line become the same as the outside, crystallinity also decreases throughout the disk.