To lessen the threats of blockading migration pathway, population isolation, and close-blood mating by check dams, previous studies suggested that removing partial dam body appropriately is one of the approaches for habitat restoration of Formosan Landlocked Salmon. After the implementation of dam remodeling for Dam #4 and Dam #3, this paper focused on the follow-up investigations on the morphological changes of observed channel. Based on the conclusions of investigation, various major findings were categorized. Based on the surveys applied at Dam No.4, the upstream detention pool of the remodeled dam has developed completely and it is expanding toward the right bank without increasing its depth for the constraints of left bank cliff and remaining dam body. The flushing chute did not change apparently between the 9th and 16th months while it expanded double in size during the following three months' heavy storm season. The original scouring pool at downstream side of dam was filled with sediment at the 9th month of remodeling and appears again by the increasing stream discharge during the next 9 months. Longitude cross section investigations indicate that 22.84% of trapped sediment beyond Dam No.4 was delivered downstream within the first 9 months and the sedimentation was slowed down to 35.7% at the 18th month for the protection of armor-layer channel boulders. Channel slope changed from its original 3.86% to 6.73% at the 9th months and stayed between 6.6% and 6.94% in the following 9 months. Affected by the factors including heavy rain in September, sediment transportation around dam site, and channel interruption by mobile equipment, the result of habitat investigation indicated that flow depth, flow area, and areas of ponds and deep flow were declined after the remodeling of No.3 Dam. Besides, the new-developed 50-meter long flush chute behind Dam No.3 destroyed several ponds. However, positive effects on channel habitat are found around dam site of Dam No.4 by its ponds and slow flow. Based on the equation developed by Abrahams & Atkinson (1995) for boulder step-pools, the steps found in the flusing chute are in the stable condition for their H/L/S values being between 1 and 2.